Non‐pharmacological management of infant and young child procedural pain

Abstract

Background

Despite evidence of the long‐term implications of unrelieved pain during infancy, it is evident that infant pain is still under‐managed and unmanaged. Inadequately managed pain in infancy, a period of exponential development, can have implications across the lifespan. Therefore, a comprehensive and systematic review of pain management strategies is integral to appropriate infant pain management. This is an update of a previously published review update in the Cochrane Database of Systematic Reviews (2015, Issue 12) of the same title.

Objectives

To assess the efficacy and adverse events of non‐pharmacological interventions for infant and child (aged up to three years) acute pain, excluding kangaroo care, sucrose, breastfeeding/breast milk, and music.

Search methods

For this update, we searched CENTRAL, MEDLINE‐Ovid platform, EMBASE‐OVID platform, PsycINFO‐OVID platform, CINAHL‐EBSCO platform and trial registration websites (ClinicalTrials.gov; International Clinical Trials Registry Platform) (March 2015 to October 2020). An update search was completed in July 2022, but studies identified at this point were added to 'Awaiting classification' for a future update. 

We also searched reference lists and contacted researchers via electronic list‐serves. 

We incorporated 76 new studies into the review. 

Selection criteria

Participants included infants from birth to three years in randomised controlled trials (RCTs) or cross‐over RCTs that had a no‐treatment control comparison. Studies were eligible for inclusion in the analysis if they compared a non‐pharmacological pain management strategy to a no‐treatment control group (15 different strategies). In addition, we also analysed studies when the unique effect of adding a non‐pharmacological pain management strategy onto another pain management strategy could be assessed (i.e. additive effects on a sweet solution, non‐nutritive sucking, or swaddling) (three strategies). The eligible control groups for these additive studies were sweet solution only, non‐nutritive sucking only, or swaddling only, respectively. Finally, we qualitatively described six interventions that met the eligibility criteria for inclusion in the review, but not in the analysis. 

Data collection and analysis

The outcomes assessed in the review were pain response (reactivity and regulation) and adverse events. The level of certainty in the evidence and risk of bias were based on the Cochrane risk of bias tool and the GRADE approach. We analysed the standardised mean difference (SMD) using the generic inverse variance method to determine effect sizes. 

Main results

We included total of 138 studies (11,058 participants), which includes an additional 76 new studies for this update. Of these 138 studies, we analysed 115 (9048 participants) and described 23 (2010 participants) qualitatively. We described qualitatively studies that could not be meta‐analysed due to being the only studies in their category or statistical reporting issues. We report the results of the 138 included studies here. An SMD effect size of 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect. The thresholds for the I² interpretation were established as follows: not important (0% to 40%); moderate heterogeneity (30% to 60%); substantial heterogeneity (50% to 90%); considerable heterogeneity (75% to 100%). The most commonly studied acute procedures were heel sticks (63 studies) and needlestick procedures for the purposes of vaccines/vitamins (35 studies). We judged most studies to have high risk of bias (103 out of 138), with the most common methodological concerns relating to blinding of personnel and outcome assessors. Pain responses were examined during two separate pain phases: pain reactivity (within the first 30 seconds after the acutely painful stimulus) and immediate pain regulation (after the first 30 seconds following the acutely painful stimulus). We report below the strategies with the strongest evidence base for each age group.

In preterm born neonates, non‐nutritive sucking may reduce pain reactivity (SMD ‐0.57, 95% confidence interval (CI) ‐1.03 to ‐0.11, moderate effect; I² = 93%, considerable heterogeneity) and improve immediate pain regulation (SMD ‐0.61, 95% CI ‐0.95 to ‐0.27, moderate effect; I² = 81%, considerable heterogeneity), based on very low‐certainty evidence. Facilitated tucking may also reduce pain reactivity (SMD ‐1.01, 95% CI ‐1.44 to ‐0.58, large effect; I² = 93%, considerable heterogeneity) and improve immediate pain regulation (SMD ‐0.59, 95% CI ‐0.92 to ‐0.26, moderate effect; I² = 87%, considerable heterogeneity); however, this is also based on very low‐certainty evidence. While swaddling likely does not reduce pain reactivity in preterm neonates (SMD ‐0.60, 95% CI ‐1.23 to 0.04, no effect; I² = 91%, considerable heterogeneity), it has been shown to possibly improve immediate pain regulation (SMD ‐1.21, 95% CI ‐2.05 to ‐0.38, large effect; I² = 89%, considerable heterogeneity), based on very low‐certainty evidence.

In full‐term born neonates, non‐nutritive sucking may reduce pain reactivity (SMD ‐1.13, 95% CI ‐1.57 to ‐0.68, large effect; I² = 82%, considerable heterogeneity) and improve immediate pain regulation (SMD ‐1.49, 95% CI ‐2.20 to ‐0.78, large effect; I² = 92%, considerable heterogeneity), based on very low‐certainty evidence. 

In full‐term born older infants, structured parent involvement was the intervention most studied. Results showed that this intervention has little to no effect in reducing pain reactivity (SMD ‐0.18, 95% CI ‐0.40 to 0.03, no effect; I² = 46%, moderate heterogeneity) or improving immediate pain regulation (SMD ‐0.09, 95% CI ‐0.40 to 0.21, no effect; I² = 74%, substantial heterogeneity), based on low‐ to moderate‐certainty evidence.

Of these five interventions most studied, only two studies observed adverse events, specifically vomiting (one preterm neonate) and desaturation (one full‐term neonate hospitalised in the NICU) following the non‐nutritive sucking intervention. The presence of considerable heterogeneity limited our confidence in the findings for certain analyses, as did the preponderance of evidence of very low to low certainty based on GRADE judgements.

Authors' conclusions

Overall, non‐nutritive sucking, facilitated tucking, and swaddling may reduce pain behaviours in preterm born neonates. Non‐nutritive sucking may also reduce pain behaviours in full‐term neonates. No interventions based on a substantial body of evidence showed promise in reducing pain behaviours in older infants. Most analyses were based on very low‐ or low‐certainty grades of evidence and none were based on high‐certainty evidence. Therefore, the lack of confidence in the evidence would require further research before we could draw a definitive conclusion.

Plain language summary

How can caregivers manage pain in infants and young children undergoing painful procedures?

Key messages

Non‐nutritive sucking (an object, such as pacifier, being placed in an infant's mouth to stimulate sucking behaviours), facilitated tucking (containing the infant using a care‐giver's hands on both head and lower limbs to maintain a 'folded‐in' position), and swaddling (wrapping the infant tightly in a blanket to prevent excessive limb movement) are among the most promising strategies that may reduce pain behaviours in preterm newborns. Non‐nutritive sucking may reduce pain behaviours in full‐term newborns. None of the strategies analysed reduced pain behaviours in older infants with sufficient evidence. Structured parent involvement (parents instructed on using strategies, such as shushing, rocking, tickling, or distraction, without being given any materials to aid them) did have a more substantial evidence base but did not have an effect on reducing pain behaviours.

Introduction of review topic

Infants and young children get exposed to several acute (lasting a short time) painful medical procedures in the first three years of life. Receiving these painful procedures without adequate pain management strategies can have negative effects on their development.

What did we want to find out?

We studied several pain management strategies (excluding kangaroo care, sucrose, breastfeeding/breast milk, and music due to existing reviews on these strategies) after acute medical procedures in preterm born newborns, full‐term born newborns, and full‐term born older infants up to the age of three, to understand how effective these strategies are at reducing pain.

What did we do?

We assessed 24 different strategies for reducing young children's pain after medical procedures using care‐giving strategies that do not require medication, such as using a pacifier, swaddling a child, and massaging a child. We compared the pain‐reducing effects of these strategies to groups receiving no pain management strategies. When possible, we also compared groups receiving one strategy to those receiving multiple strategies to see whether multiple strategies lead to more pain reduction. We looked at whether there was a difference in the impact of the interventions depending on whether the infant had just had the painful procedure (pain reactivity phase), as opposed to calming down from the peak distress (immediate pain regulation phase). 

We converted different measures of pain intensity (coded by either trained nurses or research staff) into a standard scale to help readers interpret the findings. The standard scale ranges from 0 to 21, with 0 being no pain and 21 being very severe pain. 

What did we find?

This updated Cochrane Review included 138 randomised controlled trials (trials in which participants were randomly assigned to one of two or more treatment groups) involving 11,058 participants undergoing a painful acute medical procedure. Non‐nutritive sucking, swaddling, facilitated tucking, and structured parent involvement were the four strategies most studied. 

In preterm newborns, there was evidence that non‐nutritive sucking, facilitated tucking, and swaddling may reduce pain. On the standard scale, preterm newborns receiving non‐nutritive sucking may, on average, score two points lower than preterms receiving no strategies both immediately after the painful procedure and when calming down from peak distress. Preterm newborns receiving facilitated tucking may, on average, score 3.5 points lower immediately after a painful procedure and two points lower when calming down from peak distress compared to preterms receiving no pain management strategies. While swaddling does not seem to reduce pain scores immediately after the painful procedure, swaddled preterm newborns may score, on average, four points lower than newborns receiving no strategies when calming down from distress. 

In full‐term newborns, non‐nutritive sucking may reduce pain. On the standard scale, full‐term newborns receiving non‐nutritive sucking may, on average, score four points lower immediately after a painful procedure and five points lower when calming down from peak distress compared to newborns receiving no pain management strategies.

Structured parent involvement was the strategy most studied in full‐term born older infants, but evidence showed that this strategy likely has little to no pain reduction effect in this age group. 

Adverse events were very rare across these strategies. Following non‐nutritive sucking, one preterm newborn vomited and one full‐term newborn had lower oxygen levels. No adverse events occurred following swaddling, facilitated tucking, or structured parent involvement. 

What are the limitations of the evidence?

The results of this review are based on very uncertain evidence. Many studies were too small or there were not enough studies on particular interventions to be certain about the results for our outcomes. There was also inconsistency across studies because the administration of non‐pharmacological interventions varied widely across trials in different settings. Many studies also used methods likely to introduce errors in their results. Overall, none of the analyses presented here were based on enough evidence to allow us to draw firm conclusions (i.e. high‐certainty studies from at least two independent research groups).

How up‐to‐date is the evidence?

This review is based on evidence up until October 2020. The search for studies was updated up to July 2022 and 33 eligible studies are awaiting assessment and will be incorporated in to a future update of this review.

Summary of findings

Summary of findings 1. Non‐nutritive sucking compared to no‐treatment control in preterm born neonates.

Non‐nutritive sucking compared to no‐treatment control in preterm born neonates
Patient or population: preterm born neonates Setting: inpatient Intervention: non‐nutritive sucking Comparison: no‐treatment control Countries: Italy, Saudi Arabia, Norway, Taiwan, Canada, China, USA
Outcomes	Anticipated absolute effects* (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with non‐nutritive sucking
Pain reactivity (first 30 seconds post‐procedure)	The SMD of pain reactivity in the intervention group was 0.57 lower (1.03 lower to 0.11 lower)	597 (7 RCTs)	⊕⊝⊝⊝ VERY LOWa,b	—
Immediate pain regulation (between 30 seconds and 5 minutes post‐procedure)	The SMD of immediate pain regulation in the intervention group was 0.61 lower (0.95 lower to 0.27 lower)	379 (6 RCTs)	⊕⊝⊝⊝ VERY LOWc,d	—
Adverse events	One study reported that one of the 22 participants receiving the non‐nutritive sucking intervention vomited (Gao 2018). Six studies explicitly mentioned that no adverse events occurred.	318 (7 RCTs)	⊕⊝⊝⊝ VERY LOWe,f	—
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).   Pain measures: Premature Infant Pain Profile (PIPP), Brazelton Neonatal Behavioral Assessment Scale (BNBAS), infant behavioural state; higher scores indicative of higher pain responses. CI: confidence interval; RCT: randomised controlled trial; SMD: standardised mean difference
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

^aWe downgraded the certainty of evidence by two levels for very serious study limitations as most included studies had high risk of bias (6/7 had high risk of bias overall and 1/7 had unclear risk of bias overall) because of potential issues with methods, i.e. allocation concealment, blinding of personnel and outcome assessors, and attrition.

^bWe downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I² statistic of 93%.

^cWe downgraded the certainty of evidence by two levels for very serious study limitations as most included studies had high risk of bias (5/6 had high risk of bias overall and 1/6 had unclear risk of bias overall) because of potential issues with methods, i.e. allocation concealment, blinding of personnel and outcome assessors, and attrition.

^dWe downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I² statistic of 82%.

^eWe downgraded the certainty of evidence by two levels for very serious study limitations as most included studies had high risk of bias (6/7 had high risk of bias overall and 1/6 had unclear risk of bias overall) due to various methodological concerns (e.g. blinding of personnel and outcome assessors, allocation concealment, random sequence generation, attrition bias).

^fWe downgraded the certainty of evidence by one level for inconsistency. Adverse outcomes were variable across studies, with six studies reporting no adverse outcomes while one study reported an adverse outcome.

Summary of findings 2. Non‐nutritive sucking compared to no‐treatment control in full‐term born neonates.

Non‐nutritive sucking compared to no‐treatment control in full‐term born neonates
Patient or population: full‐term born neonates Setting: inpatient Intervention: non‐nutritive sucking Comparison: no‐treatment control Countries: Italy, France, Iran, Taiwan, Turkey, USA, Hong Kong, South Korea, Vietnam
Outcomes	Anticipated absolute effects* (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with non‐nutritive sucking
Pain reactivity (first 30 seconds post‐procedure)	The SMD of pain reactivity in the intervention group was 1.13 lower (1.57 lower to 0.68 lower)	545 (8 RCTs)	⊕⊝⊝⊝ VERY LOWa,b	—
Immediate pain regulation (between 30 seconds and 5 minutes post‐procedure)	The SMD of immediate pain regulation in the intervention group was 1.49 lower (2.20 lower to 0.78 lower)	536 (9 RCTs)	⊕⊝⊝⊝ VERY LOWc,d	—
Adverse events	Thakkar 2016 reported that one participant in the treatment group and two participants in the control group desaturated during the study. The remaining studies did not report any adverse events.	507 (9 RCTs)	⊕⊝⊝⊝ VERY LOWe,f	—
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).   Pain measures: Douleur Aiguë Nouveau‐né (DAN), Neonatal Infant Pain Scale (NIPS), Premature Infant Pain Profile (PIPP), cry duration, Neonatal Facial Coding System (NFCS), Neonatal Pain, Agitation, and Sedation Scale (N‐PASS); higher scores indicative of higher pain responses. CI: confidence interval; RCT: randomised controlled trial; SMD: standardised mean difference
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

^aWe downgraded the certainty of evidence by two levels for very serious study limitations as half of the included studies had high risk of bias (4/7 had high risk of bias overall and 3/7 had unclear risk of bias overall) because of potential issues with methods, i.e. allocation concealment, blinding of personnel and outcome assessors, and attrition.

^bWe downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I² statistic of 82%.

^cWe downgraded the certainty of evidence by two levels for very serious study limitations as most included studies had high risk of bias (7/9 had high risk of bias overall and 2/9 had unclear risk of bias overall) because of potential issues with methods, i.e. allocation concealment, blinding of personnel and outcome assessors, and attrition.

^dWe downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I² statistic of 92%.

^eWe downgraded the certainty of evidence by two levels for very serious study limitations as most included studies had high risk of bias (7/9 had high risk of bias overall and 2/9 had unclear risk of bias overall) due to various methodological concerns (e.g. allocation concealment, blinding of personnel and outcome assessors, and attrition).

^fWe downgraded the certainty of evidence by one level for inconsistency. Adverse outcomes were variable across studies, with one study reporting an adverse outcome and the remaining studies not reporting any adverse events.

Summary of findings 3. Swaddling compared to no‐treatment control for preterm born neonates.

Swaddling compared to no‐treatment control for preterm born neonates
Patient or population: preterm born neonates Setting: inpatient Intervention: swaddling Comparison: no‐treatment control Countries: Taiwan, Canada, Hong Kong, Iran, Turkey
Outcomes	Anticipated absolute effects* (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with swaddling
Pain reactivity (first 30 seconds post‐procedure)	The SMD of pain reactivity in the intervention group was 0.60 lower (1.23 lower to 0.04 higher)	468 (5 RCTs)	⊕⊝⊝⊝ VERY LOWa,b	—
Immediate pain regulation (between 30 seconds and 5 minutes post‐procedure)	The SMD of immediate pain regulation in the intervention group was 1.21 lower (2.05 lower to 0.38 lower)	204 (4 RCTs)	⊕⊝⊝⊝ VERY LOWc,d,e	—
Adverse events	Five studies reported on this outcome and did not observe any adverse effects	304 (5 RCTs)	⊕⊕⊝⊝ LOWf	—
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).   Pain measures: Neonatal Infant Pain Scale (NIPS), Premature Infant Pain Profile (PIPP), Neonatal Facial Coding System (NFCS); higher scores indicative of higher pain responses. CI: confidence interval; RCT: randomised controlled trial; SMD: standardised mean difference
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

^aWe downgraded the certainty of evidence by two levels due to very serious study limitations. Risk of bias was high in 4/5 studies and unclear in 1/5 studies due to potential issues with methods, i.e. allocation concealment, blinding of personnel and outcome assessors, presence of variables confounding the treatment effect.

^bWe downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I² statistic of 91%.

^cWe downgraded the certainty of evidence by two levels due to very serious study limitations. Risk of bias was high in 3/4 studies and unclear in 1/4 studies due to potential issues with methods, i.e. blinding of personnel and outcome assessors, presence of variables confounding the treatment effect.

^dWe downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I² statistic of 89%.

^eWe downgraded the certainty of evidence one level for imprecision as n = 100 in the treatment group, which did not exceed the threshold for imprecision of 30x.

^fWe downgraded the certainty of evidence by two levels due to very serious study limitations (3/5 had high risk of bias and 2/5 had unclear risk of bias overall) because of potential issues with methods, i.e. allocation concealment, blinding of personnel, and other sources of bias.

Summary of findings 4. Facilitated tucking compared to no‐treatment control for preterm born neonates.

Facilitated tucking compared to no‐treatment control for preterm born neonates
Patient or population: preterm born neonates Setting: inpatient Intervention: facilitated tucking Comparison: no‐treatment control Countries: Finland, Brazil, USA, India, Malaysia, Iran, Taiwan, Turkey
Outcomes	Anticipated absolute effects* (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with facilitated tucking
Pain reactivity (first 30 seconds post‐procedure)	The mean pain reactivity in the intervention group was 1.01 standard deviations lower (1.44 lower to 0.58 lower)	733 (12 RCTs)	⊕⊝⊝⊝ VERY LOWa,b h	—
Immediate pain regulation (between 30 seconds and 5 minutes post‐procedure)	The mean immediate pain regulation in the intervention group was 0.59 standard deviations lower (0.92 lower to 0.26 lower)	557 (10 RCTs)	⊕⊝⊝⊝ VERY LOWc,d e	—
Adverse events	Of the 10 studies, Axelin 2006 reported that a participant got septicaemia after receiving experimental care. The other 9 studies did not observe any adverse effects.	299 (10 RCTs)	⊕⊝⊝⊝ VERY LOWf,g	—
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).   Pain measures: Neonatal Infant Pain Scale (NIPS), Premature Infant Pain Profile (PIPP), Neonatal Facial Coding System (NFCS), cry duration; higher scores indicative of higher pain responses. CI: confidence interval; RCT: randomised controlled trial; SMD: standardised mean difference
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

^aWe downgraded the certainty of evidence two levels for very serious study limitations. Risk of bias is high in 8/12 studies and unclear in 4/12 studies, due to potential issues with methods, i.e. sequence generation, allocation concealment, blinding of personnel and outcome assessors, attrition.

^bWe downgraded the certainty one level for inconsistency due to heterogeneity among studies, with an I² statistic of 93%.

^cWe downgraded the certainty of evidence two levels for very serious study limitations. Risk of bias was high in 7/10 studies and unclear in 3/10 studies. There were multiple potential issues with methods, i.e. allocation concealment, blinding of personnel and outcome assessors.

^dWe downgraded the certainty of evidence one level for imprecision as n = 278 in the treatment groups, which did not exceed the threshold for imprecision of 30x.

^eWe downgraded the certainty of evidence one level for inconsistency due to heterogeneity among studies, with an I² statistic of 87%.

^fWe downgraded the certainty of evidence two levels for very serious study limitations. Risk of bias was high in 5/10 studies and unclear in 5/10 studies. There were multiple potential issues with methods, i.e. sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition.

^g We downgraded the certainty of evidence one level for imprecision as n = 266 in the treatment groups, which did not exceed the threshold for imprecision of 30x.

^hOne study contributed two treatment arms to the pain reactivity outcome (Comaru 2009).

Summary of findings 5. Structured parent involvement compared to no‐treatment control for full‐term born older infants.

Structured parent involvement compared to no‐treatment control for full‐term born older infants
Patient or population: full‐term born older infants Setting: outpatient and inpatient Intervention: structured parent involvement Comparison: no‐treatment control Countries: Australia, USA, Canada
Outcomes	Anticipated absolute effects* (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with structured parent involvement
Pain reactivity (first 30 seconds post‐procedure)	The mean pain reactivity in the intervention group was 0.18 standard deviations lower (0.40 lower to 0.03 higher)	653 (6 RCTs)	⊕⊕⊕⊝ MODERATEa,b	—
Immediate pain regulation (between 30 seconds and 5 minutes post‐procedure)	The mean immediate pain regulation in the intervention group was 0.09 standard deviations lower (0.40 lower to 0.21 higher)	728 (4 RCTs)	⊕⊕⊝⊝ LOWa,c,d	—
Adverse events	One study reported transient skin reactions (i.e. redness, pallor) to the lidocaine cream included in their parent‐directed educational video about infant soothing (Taddio 2017). Five studies did not observe any adverse effects.	854 (6 RCTs)	⊕⊕⊝⊝ LOWe,f	—
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).   Pain measures: Neonatal Facial Coding System (NFCS), Modified Behavioral Pain Scale (MBPS), cry duration; higher scores indicative of higher pain responses. CI: confidence interval; RCT: randomised controlled trial; SMD: standardised mean difference
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

^aOne study contributed two treatment arms in each outcome (Pillai Riddell 2018).

^bWe downgraded the certainty of evidence one level for serious study limitations as 3/6 studies had high risk of bias, 1/6 had unclear risk of bias, and 2/6 had low risk of bias. However, methodological issues were due to the inability to blind personnel and outcome assessors to non‐pharmacological pain management interventions.

^cWe downgraded the certainty of evidence one level for serious study limitations as 2/4 studies had high risk of bias (because of potential issues with methods, i.e. allocation concealment, blinding of personnel and outcome assessors), while 2/4 studies had low risk of bias.

^dWe downgraded the certainty of evidence one level for inconsistency due to heterogeneity among studies, with an I² statistic of 74%.

^eWe downgraded the certainty of evidence one level for serious study limitations (3/6 studies had high risk of bias, 2/6 had low risk of bias, and 1/6 had unclear risks of bias). However, the methodological concerns were mostly due to the inability to blind personnel and outcome assessors to non‐pharmacological pain management interventions.

^fWe downgraded the certainty of evidence one level for inconsistency as there was variability across studies, with one study finding an adverse event while the others did not.

Background

This review is an update of a previously published review in the Cochrane Database of Systematic Reviews (2015, Issue 12), entitled 'Non‐pharmacological management of infant and young child procedural pain' (Pillai Riddell 2015).

Description of the condition

Despite the vigorous responses that result when an infant is subjected to a painful procedure, the premise that infants are insensitive to pain was only rejected by the general scientific community in the late 1980s (Anand 1987). Early studies suggested that infants did not possess a cortex well‐developed enough to perceive or localise pain (McGraw 1943). Moreover, initial misinterpretations of common infant pain outcomes, for example the lack of declarative memory for painful experiences during infancy (Field 1995), the muted responses of premature infants after a barrage of painful procedures (Johnston 1993), and unacceptable rates of cardiac arrest or death due to poor knowledge of infant morphokinetics during the 1950s to 1970s (Berde 2005), all perpetuated widespread neglect of infant pain treatment until the last three decades.

Neurobiological research in the 21st century now clearly denotes that infants possess the anatomical and functional requirements to perceive pain (Fitzgerald 2005; Fitzgerald 2020; Goksan 2018; Hartley 2016), and respond during tissue insult in a manner unequivocally interpretable as pain (Fabrizi 2011; Grunau 1987). Despite evidence of the long‐term implications of unrelieved pain during infancy (Anand 2000; Chau 2019; Duerden 2018; Grunau 2000; Ranger 2014), it is clearly evident that infant pain is still under‐managed and unmanaged (Anand 2017; Cruz 2016; Pillai Riddell 2013b; Waxman 2020). A comprehensive and systematic review of pain management strategies is integral to appropriate infant pain management.

Description of the intervention

Generally speaking, pain management can be subdivided into two categories: pharmacological (Barber 2004; treatments that deal with the uses, effects, and modes of action of drugs) and non‐pharmacological (any treatments (environmental, psychological/cognitive, and behavioural strategies) that do not deal with uses, effects, and modes of action of drugs).

In a précis of infant non‐pharmacological pain management strategies, pain management was defined as any strategy or technique administered to an infant in pain with the intention of lessening pain experience (Pillai Riddell 2013a). One of the important principles in infant pain management is to recognise that pain is most effectively managed by avoiding, preventing or limiting exposure to pain‐provoking stimuli (Joint Committee 2000). Accordingly, pain management during infancy must be multifaceted and integrated within every step of the decision‐making process: from deciding whether a particular procedure is warranted, to determining the safest and most efficacious pain‐relieving strategy.

There are a number of published reviews that summarise assessment and management techniques for painful procedures in neonates or infants, or both (Anand 2001; Bucsea 2019; Cignacco 2007; Cohen 2008; Johnston 2011; Kleiber 1999; Obeidat 2009; Piira 2005; Prasopkittikun 2003; Schechter 2007; Slifer 2002); however, aside from this Cochrane Review there are no other comprehensive, systematic meta‐analyses conducted on the efficacy or effectiveness of non‐pharmacological interventions for managing acute pain and distress in infants and young children up to three years of age that:

1. analyse results according to developmental subcategories within infancy (i.e. preterm‐born neonate, full‐term born neonate, full‐term born older infant); and

2. analyse type of pain phase (i.e. phase right after the painful stimulus (pain reactivity) versus those after the initial pain response period (immediate pain regulation). This dichotomy is supported by previous research (Pillai Riddell 2013d).

How the intervention might work

Different non‐pharmacological interventions will have different mechanisms of action. Therefore, cognitive, behaviour or contextual changes to the environment that are designed to manage pain would either directly or indirectly affect how the infant's nervous system processes noxious input. While the exact mechanisms of action are unclear, some examples are provided. For example, cognitive interventions (e.g. distraction) would impact an infant's ability to focus cognitive energy on processing the noxious input to the brain. Behavioural strategies (e.g. touch, massage, swaddling) would impact the noxious input on the nerve pathways to the brain. Finally, contextual interventions such as light reduction would reduce stress for the infant and thereby make the brain potentially less receptive to noxious input.

Why it is important to do this review

Unmanaged pain in infancy can have implications across the lifespan. Given the steep trajectory of infant development, both psychologically and physiologically, development within infancy is critical to acknowledge when estimating efficacy. Moreover, given the different physiological and psychological mechanisms subsuming a person’s immediate reaction to a painful stimulus (automatic or reflexive reactivity, or both) and a person’s reaction during the period of recovering from the painful insult (immediate pain regulation), it was seen as crucial to elucidate this variable to keep moving the literature on infant acute pain management forward. Therefore, this review takes a broader and more in‐depth look at the non‐pharmacological pain management literature for infant acute pain.

Objectives

To assess the efficacy and adverse events of non‐pharmacological interventions for infant and child (aged up to three years) acute pain, excluding kangaroo care, sucrose, breastfeeding/breast milk, and music.

Methods

Criteria for considering studies for this review

Types of studies

We first included randomised controlled trials (RCTs) and randomised cross‐over trials that utilised a no‐treatment control group. Randomised trials are the best design to minimise bias when evaluating the effectiveness of an intervention. We also included studies that utilised an active control group if the study design allowed for the determination of an 'additive' effect of a non‐pharmacological intervention. All studies involved the non‐pharmacological management of acute procedural pain in infants and children up to three years of age.

Due to existing work completed by other Cochrane Review authors, we excluded studies that focused on kangaroo care (Johnston 2017), sucrose (Stevens 2016a), breast milk (Shah 2012), or music (Cepeda 2013), as a pain management strategy. In addition, due to other existing Cochrane Reviews on these acute pain stimuli, we also excluded studies that examined pain management for the following types of acute pain stimuli and age group combinations: circumcision procedure for boys aged zero to three years (Brady‐Fryer 2009; Cyna 2010), and needle‐related or procedural pain in children older than three years (Birnie 2018).

Types of participants

Participants included all young children (healthy or sick) up to three years of age who were undergoing painful acute procedures. We included full‐term born and preterm‐born infants in the review.

Types of interventions

This review included care‐giving strategies that sought to lower infant pain‐related distress behaviours by cognitive, behavioural or contextual interventions. For ease of use, we grouped strategies that were considered to have a similar (albeit not identical) mechanism of action together into categories. The 24 interventions included in this review were the following.

Interventions

Contextual strategies

The context in which a painful procedure is conducted modifies behavioural and physiological expression of infant pain. Interventions that are classified in this category involved modifying the environment to lower pain reactivity and stress. They tended to involve multiple components that in combination modify the environment in which an infant experiences pain.

Sound reduction

Environmental sounds may cause stress in hospitalised infants. Therefore, this strategy involves modifying the infant's environment by reducing the amount of noise they are exposed to.

Sound addition

This strategy involves modifying the infant's environment by exposing them to a reproduction of his or her mother’s voice within the womb, designed to help simulate the soothing fetal environment.

Smell addition

This strategy involves modifying the infant's environment by exposing them to a soothing smell during the procedure, such as their mother's, breast milk, lavender, and vanilla smells.

Light reduction

This environmental strategy consists of minimising the amount of light the infant is exposed to in order to lower their pain reactivity and stress, either by direct (i.e. covering their eyes) or indirect (i.e. placing a blanket over the infants' incubator) manipulation.

Heat addition

This environmental strategy consists of warming up the site of the painful procedure (e.g. heel warming for heel sticks) using a non‐pharmacological method, such as the application of a thermal bag to the procedure site.

Cold addition

This environmental strategy consists of cooling the site of the painful procedure using a non‐pharmacological method, such as the application of an ice pack to the procedure site.

Multisensory bundle

The context in which a painful procedure is conducted modifies behavioural and physiological expressions of infant pain. This category of interventions combined multiple analgesic strategies impacting infant pain via distinct underlying mechanisms in order to lower pain behaviour. They tended to combine both direct and indirect pain management strategies (i.e. soothing smells, low lights, and noise, non‐nutritive sucking, swaddling, etc.).

Parent presence

This strategy consists of having a care‐giver present in the room during the painful procedure. Care‐givers were allowed to hold their infants' hands, but they were not given any other instructions on how to interact with their children. 

Cognitive strategies

Any intervention that is suspected to have a mechanism of action that impacts an infant’s abilities to perceive the pain experience was classified here.

Toy distraction

This distraction strategy aims to divert the infant's attention from the painful stimulus using a toy, thus impacting their ability to perceive the pain experience. 

Video distraction

This distraction strategy aims to divert the infant's attention from the painful stimulus using a video (use of an audio‐visual screen displaying two‐dimensional moving images with co‐ordinated audio output).

Behavioural strategies

These strategies involve either direct (e.g. rocking) or indirect (e.g. non‐nutritive sucking) manipulation of the infant’s body by a care‐giver. Most research on non‐pharmacological pain management strategies has been conducted within this domain.

Non‐nutritive sucking

Non‐nutritive sucking‐related strategies: an object (e.g. pacifier, non‐lactating nipple) is placed into an infant's mouth to stimulate oro‐tactile or sucking behaviours during a painful event. This may have involved other adjuvant non‐pharmacological interventions that fall under the purvey of the review (e.g. pacifier plus water was included; pacifier plus sucrose was not included unless an 'additive' study; see below).

Swallowing water

Water is administered for ingestion without an instrument that would incite extensive sucking (e.g. water administered by a dropper). This was analysed when it was an actual intervention arm, distinct to a no‐treatment control.

Rocking, holding or both

Rocking, holding or both: an infant is held or gently moved up and down or side to side (or both) by a care‐giver.

Swaddling

Swaddling is when an infant is securely wrapped in a blanket to prevent the child's limbs from moving around excessively.

Facilitated tucking

Facilitated tucking involves firmly containing the infant using a care‐giver's hands on both head and lower limbs to maintain a 'folded‐in' position. Infant may or may not be wearing clothes.

Touch/massage related

These strategies consisted of a variety of tactile‐based interventions (e.g. traditional massage, Yakson therapeutic touch, stimulating a specific pressure point on the infant's body, and applying pressure to the site of the painful procedure) designed to provide counter‐stimulation to the nociceptive input.

Therapeutic touch

This strategy consisted of the therapist using their hands to assess and re‐balance the energy field of the patient using the following steps: (1) centring their state of awareness; (2) assessing the energy field of the patient; and (3) modulating the energy field.

Vibration therapy

This strategy consisted of placing a vibration device (e.g. Buzzy) near the procedure site before the painful procedure until after the procedure ended. 

Co‐bedding

This strategy consisted of placing infants side by side in an incubator without clothing (except diaper) so they can touch each other. 

Structured parent involvement

Parents are instructed or informed of strategies that are accepted as pain‐reducing but are not given any materials to aid them. A variety of strategies may or may not be enacted such as rocking, holding, shushing, talking, rubbing, tickling, and distracting attention without toy or video.

Additive non‐pharmacological intervention on sweet solution

It has been demonstrated that sucrose is efficacious at reducing procedural pain in preterm and full‐term infants (Stevens 2016a). Furthermore, Stevens 2016a concluded that sucrose in combination with other non‐pharmacological interventions may be more efficacious than sucrose alone, with the caveat that further research is needed in this area. This group of studies allowed for the analysis of the additive effect of a variety of non‐pharmacological interventions (i.e. maternal holding, exposure to warm temperatures, non‐nutritive sucking, facilitated tucking, etc.) on a sweet solution, such as sucrose or glucose. Even though these non‐pharmacological interventions impact infant pain via different mechanisms, they were grouped together to streamline the quantitative analyses. The control group in this category received only a sweet solution, such as sucrose or glucose.

Additive non‐pharmacological intervention on non‐nutritive sucking

Previous versions of this review concluded that non‐nutritive sucking was efficacious at reducing infant pain reactivity and improving immediate pain regulation, demonstrating one of the highest effect sizes among the interventions analysed. This update seeks to explore whether the efficacy of non‐nutritive sucking can be even further enhanced by analysing the additive effect of a variety of non‐pharmacological interventions (i.e. smell addition, facilitated tucking, etc.). The control group in this category received non‐nutritive sucking only.

Additive non‐pharmacological intervention on swaddling

Previous versions of this review examined the efficacy of swaddling compared to no‐treatment control and determined that swaddling was efficacious at reducing pain reactivity and improving immediate pain regulation. This update seeks to further explore whether the efficacy of swaddling is enhanced when combined with other non‐pharmacological interventions compared to swaddling only. Therefore, the control group in this category received swaddling only.

Additive non‐pharmacological intervention on topical anaesthetic

This update seeks to explore whether the efficacy of a topical anaesthetic (e.g. LMX4 cream) is enhanced when combined with non‐pharmacological interventions compared to a topical anaesthetic only. Therefore, the control group in this category received topical anaesthetic only.

Comparators

With the exception of the three groups of analyses examining the additive effect of non‐pharmacological interventions, the comparator control group in the other analyses consisted of standard care/routine care/treatment as usual. This vastly ranged between studies as per setting protocols. While some studies specified that the infants receiving routine or standard care/treatment as usual were not exposed to any pharmacological or non‐pharmacological interventions of any kind, others defined routine/standard care as being in a cot or under a radiant warmer, parent presence, gentle touch or verbal comfort. Positioning in the control group also varied across trials (supine, side‐lying, prone, sitting up in care‐giver's lap). We specify these differences in the Characteristics of included studies tables. 

For the additive non‐pharmacological intervention on sweet solution trials, the comparator control group received sucrose.

For the additive non‐pharmacological intervention on non‐nutritive sucking trials, the comparator control group received non‐nutritive sucking.

For the additive non‐pharmacological intervention on swaddling trials, the comparator control group received swaddling.

For the additive non‐pharmacological intervention on the topical anaesthetic trial, the comparator control group received a topical anaesthetic (LMX4 cream).

Types of outcome measures

Due to the emerging verbal skills of infants and recommended assessment procedures for infant pain (Franck 2000; Stevens 2007), we made the decision to only analyse objectively measured behavioural responses to pain (i.e. Premature Infant Pain Profile (PIPP); Neonatal Infant Pain Scale (NIPS); duration of cry; Neonatal Facial Coding System (NFCS); Modified Behavioural Pain Scale (MBPS); Douleur Aiguë Nouveau‐né (DAN); N–PASS; Active/Sleep State; Behavioural Observation Pain Scale; Brazelton Neonatal Behavioral Assessment Scale (BNBAS); Face Legs Arms Cry Consolability Scale (FLACC); CRIES; infant behavioural state; Infant's Pain Questionnaire; cry frequency). Scales were mostly either coded from video or audio recording, or occasionally scored at bedside by a trained coder or a parent familiar with the child. When studies had more than one behavioural response to pain, we used pain facial expression over any other behaviour. We kept detailed logs about decisions regarding which measure (when multiple were available) was selected from an article and why. The lead review authors (RPR, OB, IS) reviewed all tables to confirm judgements made by other review authors (HG, SB, MD, CC, NR, KT, SAK, LU, DL).

We did not include physiological measures due to considerable methodological heterogeneity (including significant missing data for behavioural measures; in these situations it was rare that intention‐to‐treat analyses were done). However, if well‐established, multi‐dimensional pain measures were utilised, with behavioural and physiological indicators, and it was a reliable and valid measure, we used the total score (for example, the Premature Infant Pain Profile (PIPP); Stevens 1996) to maintain the integrity of scale. Three studies utilised measures of distress that directly followed a painful procedure, therefore they were included as a measure of pain (Corbo 2000; Ferber 2008; Whipple 2004).

Primary outcomes

The primary outcome in this review is infants' pain response. Two main principles guided the analysis of this primary outcome in the meta‐analysis for this review:

1. Under the 24 different included interventions, we first separated studies into one of three age groups due to important developmental differences across these age groups impacting their responses to the painful stimulus.

   1. Preterm born neonate ‐ infants born at 36 weeks gestation between birth up to one month post‐natal age.

   2. Full‐term born neonate ‐ term‐born infants between birth up to one month post‐natal age.

   3. Full‐term born older infant ‐ term‐born infants between one month to 36 months postnatal age.

2. Second, we further classified each study within an age group into one of two pain phases to reduce heterogeneity among included studies.

   1. Pain reactivity ‐ pain phase measured within the first 30 seconds after the acutely painful stimulus was discontinued.

   2. Immediate pain regulation ‐ pain phase measured after the first 30 seconds following the acutely painful stimulus. If multiple measurements were taken after the first 30 seconds elapsed, we utilised the measurement closest to the 30‐second time point. Our results are best understood as the immediate pain regulation response rather than regulation in a more prolonged sense.

It is important to recognise that pain measures are limited in distinguishing between infant pain and infant distress (Craig 2002). Due to the presence of an objectively painful stimulus in all studies selected for this review, however, we considered all measures of negative reactions, after the administration of a known painful stimulus, to be an indicator of an infant's pain.

After presenting the 18 analysed interventions in the above format, we then reported the six interventions that were included but described qualitatively (as they did not meet the criteria for the meta‐analysis) in the same format. 

Secondary outcomes

The secondary outcome in this review is adverse events. We screened studies for reporting of adverse events. We tracked adverse events in three categories: trial did not report on adverse events, trial reported that no adverse events occurred, or trial reported that adverse events occurred (details tracked regarding group distribution of adverse effects).

Search methods for identification of studies

We utilised both electronic, author‐driven and snowball searching techniques. We imposed no language barriers, contacted known infant pain management researchers through personal tailored emails and list‐serves, and searched trial registration sites. We included only studies with behavioural/multidimensional pain measures, and pain outcome/pain measured on the day of procedure. 

Electronic searches

In terms of published studies, we designed a unique search strategy for each of five databases in conjunction with librarians affiliated with Cochrane. For the previous update of this review we conducted the search in March 2015. For this update, we searched:

• the Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library (March 2015 to October 2020);

• MEDLINE (OVID) (March 2015 to October 2020);

• Embase (OVID) (March 2015 to October 2020);

• PsycINFO (OVID) (March 2015 to October 2020);

• Cumulative Index to Nursing and Allied Health Literature (CINAHL) (EBSCO) (March 2015 to October 2020).

See Appendix 1, Appendix 2, Appendix 3, Appendix 4, and Appendix 5 for the search strategies used.

An initial search was conducted from March 2015 to October 2020. We included all eligible trials from this search in this update. The search was then re‐run in July 2022 to include new articles published between October 2020 and July 2022. We placed articles eligible for inclusion from this second search in Characteristics of studies awaiting classification to be extracted during the next update of this review (n = 33).  

Searching other resources

We made appeals to paediatric list‐serves (Pain in Child Health (PICH), Pediatric Pain, American Psychological Association’s Division 54 (Pediatric Psychology) list‐serve) for unpublished trials relating to non‐pharmacological pain management in infants. List‐serves are electronic mailing lists that facilitate widespread distribution of information to many Internet users. We also searched trial registration websites (ClinicalTrials.gov, https://clinicaltrials.gov/; International Clinical Trials Registry Platform, https://www.who.int/ictrp/en/) and, for relevant studies, subsequently contacted the lead authors to enquire whether the study had been completed. If completed, we requested a copy of the published or unpublished results.

Finally, using relevant articles from the paediatric pain literature (i.e. articles cited in the introduction (background) section of this review), we checked reference lists for any potential articles for use in this review.

Data collection and analysis

Two independent review authors double‐coded 20% of titles and abstracts to ensure inter‐rater reliability and the agreement ranged from 97.3% to 98.7%.

Selection of studies

Eleven review authors (OB, IS, HG, SB, CC, MD, NR, KT, SAK, LU, DL) independently screened titles and abstracts of studies from literature searches for inclusion in the review. A large team was required to complete this update of the review due to the substantial number of studies included (including the 20% of studies that were double‐coded for reliability). Review authors were not blinded to authors, institutions, journals, or results. For all abstracts where relevance was questionable, the lead review authors (RPR, OB) read the full article and made the final decision regarding inclusion. RPR and BS did not review any articles that they had authored to avoid bias. 

Data extraction and management

Using the full articles, the lead review authors (RPR, OB) reviewed articles for inclusion. Subsequently, authors (OB, IS, HG, SB, CC, MD, NR, KT, SAK, LU, or DL) independently extracted data using a standard piloted form in Covidence (Covidence 2019). Two authors extracted each study and checked agreement before entry into Review Manager 5.4 (RevMan 2020). In the event of disagreement, a third review author adjudicated (RPR). We collated multiple reports of the same study, so that each study rather than each report was the unit of interest in the review. We collected characteristics of the included studies in sufficient detail to populate a table of Characteristics of included studies. We extracted the following information.

• Study design (including methods, location, sites, groups, funding sources, study author declarations of interest)

• Setting

• Participants

• Intervention(s)

• Comparator(s)

• Outcomes (including measures and time points)

• Numerical data for outcomes of interest

• Type of analysis/es presented

• Adverse events

Assessment of risk of bias in included studies

At least two review authors of the extraction team (OB, IS, HG, SB, MD, CC, NR, KT, SAK, LU, DL) scored every study included in the review for risk of bias. For articles where there was disagreement, two review authors (RPR, OB, and/or IS) made the final decision. None of the review authors conducted a risk of bias assessment for any studies for which they were an author. Upon recommendation from our Cochrane PaPaS review group, in the previous update we amended our methods to use a simplified measure of risk of bias based on the Cochrane Handbook for Systematic Reviews of Interventions 5.1.0 (Chapter 8; Higgins 2011a; Sterne 2019).

We assessed the following biases for each included study.

• Random sequence generation (checking for possible selection bias). We assessed the method used to generate the allocation sequence as:

  • low risk of bias (any truly random process, e.g. random number table; computer random number generator);

  • unclear risk of bias (insufficient detail about the method of randomisation to be able to judge the generation as 'low' or 'high' risk of bias);

  • high risk of bias (non‐random process, e.g. odd or even date of birth; hospital or clinic record number).

• Allocation concealment (checking for possible selection bias). The method used to conceal allocation to interventions prior to assignment determines whether intervention allocation could have been foreseen in advance of, or during recruitment, or changed after assignment. We assessed the methods as:

  • low risk of bias (e.g. telephone or central randomisation; consecutively numbered, sealed, opaque envelopes);

  • unclear risk of bias (insufficient detail about the method of randomisation to be able to judge the generation as 'low' or 'high' risk of bias);

  • high risk of bias (e.g. open list).

• Blinding of personnel (checking for possible performance bias). We assessed the methods used to blind study personnel from knowledge of which intervention a participant received. As the participants were infants, blinding of participants was not considered appropriate and thus we only focused on blinding of personnel. We assessed the methods as:

  • low risk of bias (study states that it was blinded and describes the method used to achieve blinding, such as identical tablets matched in appearance or smell, or a double‐dummy technique);

  • unclear risk of bias (study states that it was blinded but does not provide an adequate description of how it was achieved);

  • high risk of bias (no blinding of personnel).

• Blinding of outcome assessment (checking for possible detection bias). We assessed the methods used to blind study and outcome assessors from knowledge of which intervention a participant received. We assessed the methods as:

  • low risk of bias (study has a clear statement that outcome assessors were unaware of treatment allocation, and ideally describes how this was achieved);

  • unclear risk of bias (study states that outcome assessors were blind to treatment allocation but lacks a clear statement on how it was achieved);

  • high risk of bias (outcome assessors were not blinded).

• Incomplete outcome data (checking for possible attrition bias due to the amount, nature and handling of incomplete outcome data). We assessed the methods used to deal with incomplete data as:

  • low risk of bias (no missing outcome data; reasons for missing outcome data unlikely to be related to true outcome (for survival data, censoring unlikely to be introducing bias); missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups; we inputted missing data using 'baseline observation carried forward’ analysis);

  • unclear risk of bias (insufficient reporting of attrition/exclusions to permit a judgement of ‘low risk’ or ‘high risk’ (e.g. number randomised not stated, no reasons for missing data provided, or the study did not address this outcome));

  • high risk of bias (reason for missing outcome data is likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention groups; ‘as‐treated’ analysis done with substantial departure of the intervention received from that assigned at randomisation; potentially inappropriate application of simple imputation).

• Selective reporting (checking for reporting bias). We assessed reporting biases due to selective outcome reporting. We judged studies as:

  • low risk of bias (the study protocol is available and we reported all of the study’s prespecified (primary and secondary) outcomes that are of interest in the review in the pre‐specified way);

  • unclear risk of bias (insufficient information available to permit a judgement of ‘low risk’ or ‘high risk’);

  • high risk of bias (not all of the study’s prespecified primary outcomes have been reported; we reported one or more primary outcomes using measurements, analysis methods or subsets of the data (e.g. subscales) that were not prespecified; one or more reported primary outcomes were not prespecified (unless clear justification for their reporting is provided, such as an unexpected adverse effect); we reported one or more outcomes of interest in the review incompletely so that they cannot be entered in a meta‐analysis; the study report failed to include results for a key outcome that would be expected to have been reported for such a study).

• Other bias. Beyond the specific sources of bias mentioned above, review authors were alert for any further issues that raise concerns about the possibility of bias. This last domain is 'catch‐all' for other such sources of bias (e.g. length of washout periods in cross‐over trials). We judged studies as:

  • low risk of bias (no evidence of other potential sources of bias);

  • unclear risk of bias (insufficient information available to permit a judgement of ‘low risk’ or ‘high risk’);

  • high risk of bias (another source of bias was present).

• Overall risk of bias. As per Table 6 (Sterne 2019), if one or more domain is  "unclear" or "high" risk of bias, the overall judgement was unclear or high, respectively. Thus, we judged studies as:

  • low risk of bias (if all domains were deemed low, then the overall score was low);

  • unclear risk of bias (if at least one domain was deemed unclear, then the overall score was unclear);

  • high risk of bias (if at least one domain was deemed high, then the overall score was high).

1. Summary of overall quantitative findings.

Non‐pharmacological management of infant and young child procedural pain Patient population: preterm born neonates, full‐term born neonates, or full‐term born older infants Settings: 19 different counties (Italy, Saudi Arabia, Norway, Taiwan, Canada, China, USA, Australia, Turkey, Iran, France, Hong Kong, South Korea, Indonesia, India, Finland, Brazil, Malaysia, Vietnam) Interventions: 18 analysed interventions Comparisons: no‐treatment control or additive on either sweet solution, non‐nutritive sucking, or swaddling (as indicated in table)
	Preterm born neonates		Full‐term born neonates		Full‐term born older infants
	Pain reactivity	Immediate pain regulation	Pain reactivity	Immediate pain regulation	Pain reactivity	Immediate pain regulation
Non‐nutritive sucking‐related	SMD ‐0.57 (95% CI ‐1.03 to ‐0.11)   Very low certainty	SMD ‐0.61 (95% CI ‐0.95 to ‐0.27)   Very low certainty	SMD ‐1.13 (95% CI ‐1.57 to ‐0.68) Very low certainty	SMD ‐1.49 (95% CI ‐2.20 to ‐0.78)   Very low certainty	—	—
Swallowing water	SMD 1.10 (95% CI ‐1.56 to 3.75) Low certainty	—	—	—	—	—
Rocking or holding	—	—	SMD ‐0.09 (95% CI ‐0.61 to 0.43)   Very low certainty	SMD ‐0.84 (95% CI ‐1.15 to ‐0.53)   Very low certainty	—	—
Swaddling	SMD ‐0.60 (95% CI ‐1.23 to 0.04)   Very low certainty	SMD ‐1.21 (95% CI ‐2.05 to ‐0.38)   Very low certainty	SMD ‐0.89 (95% CI ‐1.24 to ‐0.55)   Low certainty	—	—	—
Facilitated tucking	SMD ‐1.01 (95% CI ‐1.44 to ‐0.58)   Very low certainty	SMD ‐0.59 (95% CI ‐0.92 to ‐0.26)   Very low certainty	—	—	—	—
Touch/massage‐related	SMD ‐0.92 (95% CI ‐1.60 to ‐0.25)   Very low certainty	SMD ‐1.24 (95% CI ‐1.85 to ‐0.63)   Low certainty	SMD ‐0.81 (95% CI ‐1.22 to ‐0.40)   Low certainty	SMD ‐0.92 (95% CI ‐1.57 to ‐0.28) Very low certainty	SMD ‐1.28 (95% CI ‐1.91 to ‐0.65)   Low certainty	SMD ‐1.45 (95% CI ‐2.04 to ‐0.86)   Very low certainty
Toy distraction	—	—	—	—	SMD ‐0.15 (95% CI ‐0.39 to 0.08)   Moderate certainty	SMD ‐0.08 (95% CI ‐0.50 to 0.34)     Low certainty
Structured parent involvement	—	—	—	—	SMD ‐0.18 (95% CI ‐0.40 to 0.03)   Moderate certainty	SMD ‐0.09 (95% CI ‐0.40 to 0.21)   Low certainty
Sound reduction	SMD ‐0.64 (95% CI ‐2.06 to 0.78)   Very low certainty	—	—	—	—	—
Sound addition	SMD ‐0.89 (95% CI ‐1.43 to ‐0.34)   Very low certainty	SMD ‐0.92 (95% CI ‐2.20 to 0.36)   Low certainty	—	—	—	—
Smell addition	SMD ‐0.41 (95% CI ‐0.72 to ‐0.10)   Very low certainty	SMD ‐0.57 (95% CI ‐0.88 to ‐0.27)   Very low certainty	SMD ‐0.77 (95% CI ‐1.09 to ‐0.44)   Very low certainty	SMD ‐0.41 (95% CI ‐0.80 to ‐0.01)   Moderate certainty	—	—
Light reduction	SMD ‐0.71 (95% CI ‐1.08 to ‐0.34)   Moderate certainty	SMD ‐1.16 (95% CI ‐1.53 to ‐0.78)   Moderate certainty	—	—	—	—
Heat addition	—	—	SMD ‐0.12 (95% CI ‐0.42 to 0.17)   Low certainty	—	—	—
Cold addition	—	—	SMD ‐0.85 (95% CI ‐1.48 to ‐0.23)   Low certainty	—	—	—
Multisensory bundle	SMD ‐2.70 (95% CI ‐4.16 to ‐1.23)   Very low certainty	SMD ‐1.49 (95% CI ‐2.92 to ‐0.06)   Very low certainty	SMD 0.00  (95% CI ‐0.29 to 0.30) Moderate certainty	—	—	—
	Additive studies
Additive non‐pharmacological intervention on sweet solutions	SMD ‐1.10 (95% CI ‐1.96 to ‐0.25)   Low certainty	SMD ‐0.90 (95% CI ‐2.00 to 0.20)   Low certainty	SMD ‐1.26 (95% CI ‐2.45 to ‐0.07)   Low certainty	SMD ‐0.53 (95% CI ‐0.93 to ‐0.13)   Low certainty	SMD ‐0.14 (95% CI ‐0.39 to 0.1)   Moderate certainty	—
Additive non‐pharmacological intervention on non‐nutritive sucking (NNS)	SMD ‐1.27 (95% CI ‐2.21 to ‐0.33)   Very low certainty	SMD ‐1.32 (95% CI ‐2.59 to ‐0.05)   Very low certainty	—	—	—	—
Additive non‐pharmacological intervention on swaddling	—	—	SMD ‐0.98 (95% CI ‐2.40 to 0.44)   Very low certainty	—	—	—

Very low certainty in the evidence (GRADE criteria)

Low certainty in the evidence (GRADE criteria)

Moderate certainty in the evidence (GRADE criteria)

High certainty in the evidence (GRADE criteria)

Measures of treatment effect

We treated all of the outcome data for the included studies as continuous. In collaboration with Cochrane statistical consultants, we calculated the standardised mean difference (SMD) and the standard error of the mean difference (SE) separately for analysed between‐group and cross‐over trials, consistent with the procedures outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021). 

We standardised all the scores using the standardised mean difference (SMD) method recommended by Cochrane. Since all the outcomes were behaviourally based and standardised but used different types of measurements, we examined them together as either the outcome variable called 'Pain reactivity' or 'Immediate pain regulation'. As such, we treated our outcome measure the same for each intervention and it encapsulated all the above measures of pain‐related distress.

We displayed standardised mean differences (SMDs) using a random‐effects model in the results below with the 95% confidence intervals included in brackets. All participants included in the analyses were randomised and met the inclusion criteria outlined earlier. 

In collaboration with Cochrane statistical consultants, for the analyses highlighted in the Plain Language Summary, we converted the pooled SMD into a score difference on the well‐established Premature Infant Pain Profile (PIPP) using a representative SD of 3.3 for easier interpretation. 

We qualitatively compared studies that examined a relevant non‐pharmacological treatment but did not meet the inclusion criteria for the meta‐analysis. For these studies, we reported the study characteristics, risk of bias, authors' conclusions, overall certainty of the evidence based on the GRADE system, and adverse events. 

In terms of interpretation of the effect sizes, 0.2 represents a small effect, 0.5 a moderate effect, and 0.8 a large effect (Cohen 1988). 

Unit of analysis issues

For parallel design RCTs with two or more treatment arms in the same meta‐analysis, we split the control group to avoid double‐counting. 

For cross‐over RCTs, we used a correlation coefficient of 0.375, which was obtained from a review of the literature (Liaw 2012). We used this correlation coefficient in the analysis of cross‐over trials with the important caveat that this review examines different interventions that vary in their effectiveness across age groups and pain phases. We conducted sensitivity analyses to investigate the robustness of our conclusions. This approach was approved by Cochrane statistical support. Calculations are available from the authors upon request. For cross‐over trials, since all participants received both interventions as part of the design, we entered the total number of participants in both the treatment and control group. However, it should be noted that the correlation has been accounted for in the effect estimate and SE so that the data entered do not have a unit of analysis issue (Higgins 2022). 

Dealing with missing data

For every study that met the inclusion criteria, we extracted means and standard deviations from the papers. If these were not provided, we first contacted study authors to obtain the missing data. If we were unable to obtain the data, we attempted to derive the numbers needed from data provided in the article using the procedures outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021), or published in peer‐reviewed journals (Wan 2014). As an additional level of quality assurance for the review, whenever data were interpolated from a study, a sub‐team of authors (OB, IS, RPR) also conducted a conceptual double‐check ensuring confidence intervals (CIs) from interpolated data were always in line with the findings from the original study authors. If not enough information was provided to calculate the means and standard deviations, we described those studies qualitatively. 

Assessment of heterogeneity

Given our primary interest in the impact of the heterogeneity (not the presence of heterogeneity), we utilised the I² statistic (Higgins 2003). The thresholds for the I² interpretation are established as follows: not important (0% to 40%); moderate heterogeneity (30% to 60%); substantial heterogeneity (50% to 90%); considerable heterogeneity (75% to 100%), from Chapter 10 in the 2011 Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011a; Higgins 2021). We provide more details regarding this approach in the Subgroup analysis and investigation of heterogeneity section below.

Assessment of reporting biases

In order to help overcome publication bias, we imposed no language barriers, contacted known infant pain management researchers through personal emails and list‐serves, and searched trial registration sites.

Data synthesis

We organised results according to intervention, then age group, then pain phase. For example, we conducted an analysis to determine if non‐nutritive sucking was efficacious for preterm infants’ pain reactivity.

Two main principles guided the meta‐analysis of the data collected for this review. Under the 18 different meta‐analysed interventions, we first separated studies into three age‐specific categories (preterm born neonates, full‐term born neonates, and full‐term born older infants). Second, each study with an age subcategory was further divided into one of two pain phases to reduce heterogeneity among included studies (pain reactivity, immediate pain regulation).

We pooled the results from individual intervention studies using the random‐effects model in Review Manager 5.4 (RevMan 2020), weighted by study sample size, and reporting SMDs with 95% confidence intervals (CIs), and standard error (SE) of the pooled treatment effect.

Subgroup analysis and investigation of heterogeneity

Although we utilised a statistical analysis plan to limit heterogeneity among studies (i.e. treatment x age x pain response analyses), the existence of heterogeneity between studies was inevitable. Given our primary interest in the impact of the heterogeneity (not the presence of heterogeneity), we utilised the I² statistic (Higgins 2003). In cases where considerable heterogeneity (I ²> 75% as per the thresholds in the Assessment of heterogeneity section above) was found, when possible (i.e. more than two studies were analysed), we re‐ran the analyses without the source of heterogeneity (noted using the SMD) and compared to the original findings. We incorporated these data in our GRADE analyses whenever possible and noted it in the footnotes of our individual summary of findings tables.

Sensitivity analysis

We investigated factors that may have affected our overall results from individual studies using sensitivity analyses. For each pooled result, we conducted the following sensitivity analyses using Review Manager 5.4 software (RevMan 2020).

1. We compared each pooled result to the individual studies that contributed to the overall pooled result to determine if any studies were more influential than others. We re‐ran the analyses (i.e. the SMD statistic) without significantly influential studies when there were two or more studies left to contribute to an overall pooled result. This was conducted in cases of considerable heterogeneity (I²> 75%) or poor study quality/intervention integrity. For every intervention analysis, we examined the studies that contributed to the overall pooled result in greater detail when they did not agree with the overall pooled result. We examined studies for methodological differences to help offer potential reasons why there was disagreement between the studies. If conducted, we included the results of these comparisons in the 'Summary of treatment effects' section at the end of each analysis.

2. Despite our attempt to conduct analyses utilising homogeneous groups of pain management interventions, age groups, and pain response phases, we found substantial heterogeneity (I²> 75%)among many overall treatment effects. We still presented pooled results, and we added attempts to explain heterogeneity among treatment effects to the 'Summary of treatment effects' section at the end of each treatment analysis. 

Summary of findings and assessment of the certainty of the evidence

Two review authors (OB, IS) independently rated the certainty of the body of evidence for the outcomes. We used the GRADE system to rank the certainty of the evidence using the GRADEprofiler Guideline Development Tool software (GRADEpro GDT), and the guidelines provided in the CochraneHandbook for Systematic Reviews of Interventions (Chapter 14; Higgins 2019) and GRADEpro Handbook (Schünemann 2013) as the criteria for assessing the integrity of intervention. The GRADE tables follow the analysis structure of intervention x age x time point.

The GRADE approach uses five considerations (study limitations (risk of bias), unexplained heterogeneity and inconsistency of effect, imprecision, indirectness, and publication bias) to assess the certainty of the body of evidence for each outcome. The GRADE system uses the following criteria for assigning grade of evidence.

• High: we are very confident that the true effect lies close to that of the estimate of the effect

• Moderate: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different

• Low: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect

• Very low: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect

The GRADE system considers study design as a marker of quality. Randomised controlled trials are considered to be high‐quality evidence and can be downgraded for important limitations.

Factors that may decrease the certainty level of a body of evidence are as follows.

• Serious or very serious study limitations (based on methodological limitations addressed in the risk of bias assessment)

• Important or serious inconsistency of results (based on analysis of heterogeneity outlined above)

• Some or major indirectness of evidence

• Serious or very serious imprecision (based on a total treatment number ('n') that exceeds 30x where x is the number of studies)

• Probability of publication bias

We included five summary of findings tables to present the interventions based on the most substantial body of literature in a transparent and simple tabular format. In particular, we included key information concerning the certainty of the evidence, the magnitude of the effect of the interventions examined, and the sum of available data on the primary outcomes (pain reactivity, immediate pain regulation). We also reported on adverse events qualitatively. In addition, we created one overarching summary of findings table to summarise the entirety of the meta‐analysis findings in one place (see Table 6 'Summary of overall quantitative findings' table).

Results

Description of studies

See the Characteristics of included studies and Characteristics of excluded studies tables.

Results of the search

We included a total of 62 studies from the 2015 update of the review in this current update.

In this current update, after removing duplicates, we screened 11,275 potential records from March 2015 to July 2022 for eligibility, resulting in a final selection of 76 new studies added to the review (an update search from October 2020 to July 2022 included studies in Characteristics of studies awaiting classification). Along with the 62 studies from the previous review, a total of 138 studies (11,058 participants) were included in the qualitative synthesis. Of these 138 included studies, 115 were analysed in the quantitative meta‐analysis (a breakdown of quantitative + qualitative studies from the 2015 version and this update can be found in Figure 1). We excluded studies from the meta‐analysis under three conditions (described below). 

1.

Study flow diagram

The 2015 update also described qualitatively 22 papers that did not meet the RCT inclusion criteria (i.e. due to not having a control group/having an active control group) (Aguirre 2008; Bueno 2010; Campos 1989; Cignacco 2008; Cong 2009; Diego 2009; Felt 2000; Goubet 2003; Goubet 2007; Grunau 2004; Huang 2004; Ipp 2009; Johnston 2008a; Johnston 2009; Lima 2013; Ludington‐Hoe 2005; Morelius 2009; Okan 2010; Rattaz 2005; Singh 2012; Vivancos 2010; Weissman 2009). However, in an attempt to streamline this review as more relevant RCTs are included with each update, the current update now excludes these articles.

Included studies

In the case of studies where only some of the treatment arms were included in the review, we counted only the participants in the treatment arms that met the inclusion criteria in the descriptions below. Figure 1 demonstrates the flow of studies included in the review.

Of the 138 included studies (11,058 participants), 23 studies that were not meta‐analysed, but we felt were relevant, were described qualitatively for the purpose of the review. Studies were included, but described qualitatively if they 1) were the only studies in their category (e.g. if the group "rocking/holding X preterm X pain reactivity" was composed of only one study, we would not include this group in the meta‐analyses) (16 studies: Allen 1996; Badiee 2014; Bahorski 2015; Cetinkaya 2020; Cohen 2002; Cohen 2006; Curtis 2007; Erdogan 2020; Harrington 2012; Ipp 2004; Johnston 2013; Kozub 2001; Kucukoglu 2015; Secil 2014; Shaw 1982; Taddio 2017a); 2) did not provide means/SDs (four studies: Azarmnejad 2015; Benjamin 2016; Petrillo 2010; Ti 2015); 3) reported inconsistent statistics throughout text (three studies: Dabas 2019; Karaca 2016; Khasanah 2019). In this third instance, if we were not able to get a hold of study authors to verify the correct statistics, which precluded us from analysing these studies, we instead described their key conclusions qualitatively.

To enhance clarity throughout the review, we refer to studies included in the meta‐analyse as 'analysed' studies. Studies included but not meta‐analysed are referred to as 'included qualitatively'.

Of the 138 studies, 11 studies with multiple treatment arms had treatment arms included in both the quantitative and qualitative analyses (Ayazi 2017; Bauchner 1996; Carbajal 1999; Elserafy 2009; Erkut 2017; Harrington 2012; Khosravan 2018; Mitchell 2016; Najafi 2019; Shu 2014; Thakkar 2016). When describing these studies, we specified which treatment arm we are referring to (i.e. 'analysed' versus 'included qualitatively'). 

We provide the study characteristics of all 138 included studies below. 

Study design characteristics

In the 108 between‐group studies, there were 5597 participants in the treatment conditions and 4479 in the control conditions, for a total of 10,076 participants. In the 30 cross‐over studies, 982 participants were exposed to both the treatment and control conditions (these participants were only counted once towards the total number of participants). In total, 11,058 participants were included in the review. Of these, we analysed 9048 and included but discussed qualitatively 2010. 

• Cross‐over designs (30 studies; Alinejad‐Naeini 2014; Axelin 2006; Axelin 2009; Bellieni 2001; Bo 2000; Catelin 2005; Chik 2012; Chik 2017; Chuang 2019; Comaru 2009; Corbo 2000; Corff 1995; Davari 2018; Elserafy 2009; Fatollahzade 2020; Fearon 1997; Herrington 2007; Herrington 2014; Hill 2005; Jain 2006; Johnston 2007a; Kozub 2001; Kristoffersen 2011; Liaw 2012; Ranjbar 2020; Sizun 2002; Stevens 1999; Sundaram 2013; Ward‐Larson 2004; Yu 2017).

• Between‐groups designs (108 studies; Abbasoglu 2015; Akcan 2016; Alemdar 2017; Alemdar 2017a; Alemdar 2018; Alemdar 2020; Allen 1996; Angeles 2020; Apaydin Cirik 2020; Ayazi 2017; Aydin 2019; Azarmnejad 2015; Badiee 2014; Bahorski 2015; Basiri‐Moghadam 2014; Bauchner 1996; Baudesson 2017; Behnam Vashani 2015; Beiranvand 2020; Bellieni 2002; Bembich 2018; Benjamin 2016; Blass 1999; Bustos 2008; Caglar 2017; Campbell‐Yeo 2012; Campos 1994; Carbajal 1999; Carbajal 2003; Cetinkaya 2020; Chang 2020; Chen 2017; Chen 2019; Chirico 2017; Chou 2012; Cohen 2002; Cohen 2006; Cramer‐Berness 2005a; Cramer‐Berness 2005b; Curtis 2007; Dabas 2019; Dehghani 2019; Dezhdar 2016; Dur 2020; Efendi 2018; Erdogan 2020; Erkut 2017; Esfahani 2013; Fitri 2020; Gao 2018; Göl 2017; Gormally 2001; Gray 2015; Greenberg 2002; Harrington 2012; Hashemi 2016; Hillgrove Stuart 2013; Ho 2016; Hogan 2014; Im 2008; Ipp 2004; Jebreili 2015; Johnston 1997; Johnston 2013; Jose 2012; Kahraman 2020; KarabiyikOgurlu 2020; Karaca 2016; Khasanah 2019; Khosravan 2018; Koç 2015; Kucukoglu 2015; Leng 2016; Liaw 2010; Liaw 2011; Liaw 2013; Liu 2010; Lopez 2015; Malakian 2017; McGinnis 2016; Mir 2018; Mirzarahimi 2013; Mitchell 2016; Morrow 2010; Najafi 2019; Ozkan 2019; Peng 2018; Perroteau 2018; Petrillo 2010; Pillai Riddell 2018; Razaghi 2020; Ren 2019; Roshanray 2020; Sadathosseini 2013; Secil 2014; Shaw 1982; Shu 2014; Taddio 2014; Taddio 2015; Taddio 2017; Taddio 2017a; Taplak 2020; Thakkar 2016; Ti 2015; Vu‐Ngoc 2020; Whipple 2004; Yilmaz 2010; Yilmaz 2020).

Study nationality characteristics

The 138 included studies were conducted by authors in 19 different counties (Italy, Saudi Arabia, Norway, Taiwan, Canada, China, USA, Australia, Turkey, Iran, France, Hong Kong, South Korea, Indonesia, India, Finland, Brazil, Malaysia, Vietnam).

Description of study treatment arms

In total, the 138 included studies contributed 255 separate treatment arms (each of which was compared to a control group). Of the 255 total included treatment arms, we analysed 215 while the remaining 40 we described qualitatively. When we examined the additive effects of one treatment on another, it was considered as having one treatment arm. In analyses where more than one treatment arm was analysed, we split the control group n to avoid the bias associated with 'double‐counting'.

One treatment arm: 100 studies provided one treatment arm. Of these 100 studies, 30 provided data for both pain reactivity and immediate pain regulation, providing 60 treatment arms (Alemdar 2020; Axelin 2006; Ayazi 2017; Behnam Vashani 2015; Benjamin 2016; Beiranvand 2020; Campbell‐Yeo 2012; Catelin 2005; Chuang 2019; Corbo 2000; Davari 2018; Dezhdar 2016; Erdogan 2020; Erkut 2017; Göl 2017; Gormally 2001; Ho 2016; Johnston 2013; Karaca 2016; Liaw 2010; Liu 2010; McGinnis 2016; Peng 2018; Ranjbar 2020; Secil 2014; Sundaram 2013; Taddio 2015; Taddio 2017; Taddio 2017a; Yilmaz 2010). One study provided data for two different samples of older infants (Pillai Riddell 2018 (six‐month older infant); Pillai Riddell 2018 (18‐month older infant) for both reactivity and regulation), providing four treatment arms. Two studies provided two comparisons as the results were reported separately by order of treatment, contributing four treatment arms (Chik 2012; Comaru 2009). Sixty‐seven studies provided data on either pain reactivity or immediate pain regulation, providing 67 treatment arms (Abbasoglu 2015; Alinejad‐Naeini 2014; Angeles 2020; Axelin 2009; Aydin 2019; Azarmnejad 2015; Badiee 2014; Bahorski 2015; Basiri‐Moghadam 2014; Baudesson 2017; Bellieni 2001; Bellieni 2002; Bembich 2018; Blass 1999; Bo 2000; Bustos 2008; Caglar 2017; Carbajal 2003; Cetinkaya 2020; Chang 2020; Chen 2017; Chen 2019; Chik 2017; Chirico 2017; Chou 2012; Cohen 2002; Cohen 2006; Corff 1995; Curtis 2007; Dabas 2019; Dehghani 2019; Efendi 2018; Esfahani 2013; Fatollahzade 2020; Fearon 1997; Fitri 2020; Gray 2015; Greenberg 2002; Hashemi 2016; Herrington 2007; Herrington 2014; Hill 2005; Hogan 2014; Ipp 2004; Jain 2006; Johnston 1997; Johnston 2007a; Jose 2012; KarabiyikOgurlu 2020; Khasanah 2019; Koç 2015; Kozub 2001; Kucukoglu 2015; Liaw 2011; Lopez 2015; Malakian 2017; Morrow 2010; Perroteau 2018; Razaghi 2020; Shaw 1982; Sizun 2002; Taddio 2014; Ti 2015; Vu‐Ngoc 2020; Ward‐Larson 2004; Whipple 2004; Yu 2017).

Two treatment arms: 31 studies provided two treatment arms. Twenty of these studies provided data on either pain reactivity or immediate pain regulation, contributing 40 treatment arms (Allen 1996; Bauchner 1996; Campos 1994; Carbajal 1999; Cramer‐Berness 2005a; Cramer‐Berness 2005b; Harrington 2012; Im 2008; Khosravan 2018; Leng 2016; Liaw 2012; Mir 2018; Mirzarahimi 2013; Petrillo 2010; Ren 2019; Sadathosseini 2013;  Shu 2014; Stevens 1999; Thakkar 2016; Yilmaz 2020). Eleven of these 31 studies provided two arms for both pain reactivity and immediate pain regulation comparisons, with 44 treatment arms total (Alemdar 2017; Apaydin Cirik 2020; Dur 2020; Gao 2018; Hillgrove Stuart 2013; Jebreili 2015; Liaw 2013; Mitchell 2016; Najafi 2019; Ozkan 2019; Roshanray 2020).

Three or more treatment arms: Seven studies contributed three treatment arms. Four studies had comparisons of both pain reactivity and immediate pain regulation, each with three treatment arms, resulting in 24 treatment arms in total (Akcan 2016; Alemdar 2017a; Alemdar 2018; Taplak 2020). Two studies provided data on pain reactivity only, contributing six treatment‐arm comparisons (Kahraman 2020; Kristoffersen 2011). One study provided data on both pain reactivity and immediate pain regulation for three distinct interventions, contributing six treatment arms (Elserafy 2009).

Description of pain responses (pain reactivity versus immediate pain regulation)

A breakdown of which 138 included studies contributed to overall findings for pain reactivity (first 30 seconds after the pain stimulus) and immediate pain regulation (occurring between 30 seconds and five minutes post painful stimulus) is as follows.

Sixty‐two studies reported pain reactivity outcomes, which was zero to 30 seconds after the painful stimulus (Abbasoglu 2015; Alinejad‐Naeini 2014; Angeles 2020; Axelin 2009; Aydin 2019; Badiee 2014; Basiri‐Moghadam 2014; Baudesson 2017; Bellieni 2001; Bellieni 2002; Bembich 2018; Benjamin 2016; Bustos 2008; Carbajal 1999; Carbajal 2003; Chang 2020; Chen 2017; Chen 2019; Chik 2012; Chik 2017; Chirico 2017; Chou 2012; Cohen 2002; Comaru 2009; Cramer‐Berness 2005a; Cramer‐Berness 2005b; Dabas 2019; Esfahani 2013; Fatollahzade 2020; Fitri 2020; Hashemi 2016; Hill 2005; Hogan 2014; Ipp 2004; Johnston 1997; Johnston 2007a; Kahraman 2020; KarabiyikOgurlu 2020; Khasanah 2019; Khosravan 2018; Koç 2015; Kozub 2001; Kristoffersen 2011; Kucukoglu 2015; Lopez 2015; Malakian 2017; Mir 2018; Mirzarahimi 2013; Morrow 2010; Perroteau 2018; Petrillo 2010; Ranjbar 2020; Razaghi 2020; Shaw 1982; Shu 2014; Sizun 2002; Stevens 1999; Taddio 2014; Thakkar 2016; Ward‐Larson 2004; Yilmaz 2020; Yu 2017).

Twenty‐eight studies reported immediate pain regulation outcomes, which was the pain measurement closest to 30 seconds after the last painful stimulus (Allen 1996; Bahorski 2015; Bauchner 1996; Beiranvand 2020; Blass 1999; Bo 2000; Campos 1994; Cohen 2006; Corff 1995; Curtis 2007; Dehghani 2019; Efendi 2018; Fearon 1997; Gray 2015; Greenberg 2002; Harrington 2012; Herrington 2007; Herrington 2014; Im 2008; Jain 2006; Jose 2012; Leng 2016; Liaw 2011; Liaw 2012; Ren 2019; Sadathosseini 2013; Vu‐Ngoc 2020; Whipple 2004).

Forty‐eight studies reported pain outcomes for both pain reactivity and immediate pain regulation (Akcan 2016; Alemdar 2017; Alemdar 2017a; Alemdar 2018; Alemdar 2020; Apaydin Cirik 2020; Axelin 2006; Ayazi 2017; Azarmnejad 2015; Behnam Vashani 2015; Caglar 2017; Campbell‐Yeo 2012; Catelin 2005; Cetinkaya 2020; Chuang 2019; Corbo 2000; Davari 2018; Dezhdar 2016; Dur 2020; Elserafy 2009; Erdogan 2020; Erkut 2017; Gao 2018; Göl 2017; Gormally 2001; Hillgrove Stuart 2013; Ho 2016; Jebreili 2015; Johnston 2013; Karaca 2016; Liaw 2010; Liaw 2013; Liu 2010; McGinnis 2016; Mitchell 2016; Najafi 2019; Ozkan 2019; Peng 2018; Pillai Riddell 2018; Roshanray 2020; Secil 2014; Sundaram 2013; Taddio 2015; Taddio 2017; Taddio 2017a; Taplak 2020; Ti 2015; Yilmaz 2010).

Participants

Age groupings

Of all the 138 included studies, there were 61 studies that had preterm infants as participants (i.e. born at 36 weeks gestation or less). Forty‐three had participants in the neonate age category (i.e. infants born at 37 weeks until one month of age). Thirty‐three papers included older infants (i.e. over one month to 36 months of age) as participants. One paper included more than one age group. 

Preterm‐born neonate (61 studies; Abbasoglu 2015; Alemdar 2017; Alemdar 2017a; Alemdar 2018; Alemdar 2020; Alinejad‐Naeini 2014; Angeles 2020; Apaydin Cirik 2020; Axelin 2006; Axelin 2009; Ayazi 2017; Badiee 2014; Baudesson 2017; Behnam Vashani 2015; Bellieni 2001; Campbell‐Yeo 2012; Catelin 2005; Chen 2017; Chen 2019; Chik 2012; Chirico 2017; Chuang 2019; Comaru 2009; Corbo 2000; Corff 1995; Davari 2018; Dehghani 2019; Dezhdar 2016; Dur 2020; Efendi 2018; Elserafy 2009; Fatollahzade 2020; Fearon 1997; Fitri 2020; Gao 2018; Herrington 2007; Herrington 2014; Hill 2005; Ho 2016; Jain 2006; Jebreili 2015; Johnston 1997; Johnston 2007a; Johnston 2013; Kahraman 2020; Kristoffersen 2011; Liaw 2010; Liaw 2012; Liaw 2013; Lopez 2015; Peng 2018; Perroteau 2018; Ranjbar 2020; Ren 2019; Shu 2014; Sizun 2002; Stevens 1999; Sundaram 2013; Taplak 2020; Ward‐Larson 2004; Whipple 2004).

Full‐term born neonate (43 studies; Akcan 2016; Aydin 2019; Azarmnejad 2015; Bellieni 2002; Bembich 2018; Blass 1999; Bo 2000; Caglar 2017; Campos 1994; Carbajal 1999; Carbajal 2003; Cetinkaya 2020; Chang 2020; Chik 2017; Chou 2012; Dabas 2019; Erkut 2017; Gormally 2001; Gray 2015; Greenberg 2002; Hashemi 2016; Im 2008; KarabiyikOgurlu 2020; Kucukoglu 2015; Leng 2016; Liu 2010; Malakian 2017; McGinnis 2016; Mir 2018; Mirzarahimi 2013; Mitchell 2016; Morrow 2010; Najafi 2019; Ozkan 2019; Razaghi 2020; Roshanray 2020; Sadathosseini 2013; Thakkar 2016; Ti 2015; Vu‐Ngoc 2020; Yilmaz 2010; Yilmaz 2020; Yu 2017).

Full‐term born older infant (33 studies; Allen 1996; Bahorski 2015; Basiri‐Moghadam 2014; Bauchner 1996; Beiranvand 2020; Benjamin 2016; Bustos 2008; Cohen 2002; Cohen 2006; Cramer‐Berness 2005a; Cramer‐Berness 2005b; Curtis 2007; Erdogan 2020; Esfahani 2013; Göl 2017; Harrington 2012; Hillgrove Stuart 2013; Hogan 2014; Ipp 2004; Jose 2012; Karaca 2016; Khasanah 2019; Khosravan 2018; Koç 2015; Kozub 2001; Liaw 2011; Pillai Riddell 2018; Secil 2014; Shaw 1982; Taddio 2014; Taddio 2015; Taddio 2017; Taddio 2017a).

Two age groups: One study included a group of full‐term born neonates and preterm born neonates whereby both age groups were analysed separately. 

It should be noted that some studies combined two age groups into a single treatment group. We categorised these studies as one age group and only included them once in the results. Catelin 2005 had 66% preterm neonate participants and 33% full‐term neonate participants. We categorised this study as a preterm study. One study had both preterm and full‐term neonates, where the average age was 35 weeks (Fitri 2020). We categorised this study as a preterm study. Another study had 46.7% preterm neonate participants and 53.3% full‐term neonate participants (Cetinkaya 2020). We categorised this study as a full‐term study.

Diagnostic status

The diagnostic status of the infants in the 138 included studies is as follows:

Healthy infants born full‐term (74 studies; Akcan 2016; Allen 1996; Aydin 2019; Azarmnejad 2015; Bahorski 2015; Basiri‐Moghadam 2014; Bauchner 1996; Beiranvand 2020; Bellieni 2002; Bembich 2018; Benjamin 2016; Blass 1999; Bustos 2008; Caglar 2017; Campos 1994; Carbajal 1999; Carbajal 2003; Cetinkaya 2020; Chang 2020; Chik 2017; Chou 2012; Cohen 2002; Cohen 2006; Cramer‐Berness 2005a; Cramer‐Berness 2005b; Curtis 2007; Dabas 2019; Erkut 2017; Esfahani 2013; Gormally 2001; Gray 2015; Greenberg 2002; Göl 2017; Harrington 2012; Hashemi 2016; Hillgrove Stuart 2013; Hogan 2014; Im 2008; Ipp 2004; Jose 2012; KarabiyikOgurlu 2020; Karaca 2016; Khasanah 2019; Khosravan 2018; Koç 2015; Kucukoglu 2015; Leng 2016; Liaw 2011; Liu 2010; Malakian 2017; McGinnis 2016; Mir 2018; Mirzarahimi 2013; Mitchell 2016; Morrow 2010; Najafi 2019; Ozkan 2019; Petrillo 2010; Pillai Riddell 2018; Razaghi 2020; Roshanray 2020; Sadathosseini 2013; Secil 2014; Shaw 1982; Taddio 2014; Taddio 2015; Taddio 2017; Taddio 2017a; Thakkar 2016; Ti 2015; Vu‐Ngoc 2020; Yilmaz 2010; Yilmaz 2020; Yu 2017).

Hospitalised infants born preterm (61 studies; Abbasoglu 2015; Alemdar 2017; Alemdar 2017a; Alemdar 2018; Alemdar 2020; Alinejad‐Naeini 2014; Angeles 2020; Apaydin Cirik 2020; Axelin 2006; Axelin 2009; Ayazi 2017; Badiee 2014; Baudesson 2017; Behnam Vashani 2015; Bellieni 2001; Campbell‐Yeo 2012; Catelin 2005; Chen 2017; Chen 2019; Chik 2012; Chirico 2017; Chuang 2019; Comaru 2009; Corbo 2000; Corff 1995; Davari 2018; Dehghani 2019; Dezhdar 2016; Dur 2020; Efendi 2018; Elserafy 2009; Fatollahzade 2020; Fearon 1997; Fitri 2020; Gao 2018; Herrington 2007; Herrington 2014; Hill 2005; Ho 2016; Jain 2006; Jebreili 2015; Johnston 1997; Johnston 2007a; Johnston 2013; Kahraman 2020; Kristoffersen 2011; Liaw 2010; Liaw 2012; Liaw 2013; Lopez 2015; Peng 2018; Perroteau 2018; Ranjbar 2020; Ren 2019; Shu 2014; Sizun 2002; Stevens 1999; Sundaram 2013; Taplak 2020; Ward‐Larson 2004; Whipple 2004).

Infants born full‐term, being monitored or treated for health complications (three studies; Bo 2000; Erdogan 2020; Kozub 2001).

Types of painful procedures

Given that research in the area of infant pain management began in the late 1980s, we selected a broad mandate of 'procedural pain' rather than any particular type of procedure. For a study to be included in the search, the procedure had to be described by the author as painful or causing pain. As diaper changes and weighing procedures have been deemed painful in preterm neonates only, we included these procedures for this age group. In order to provide general parameters regarding procedures that are covered by the review, sample procedures are provided. Based on two comprehensive references that outline painful procedures in either neonates or older children (Anand 2001; Birnie 2018), the following non‐exhaustive list is provided as a sample of procedures that fell under the umbrella of this review (see Table 7). We derived definitions from two online medical encyclopaedic reference sources (i.e. MEDLINE Plus Medical Encyclopaedia: www.nlm.nih.gov/medlineplus/mplusdictionary.html; the Merck Manual of Diagnosis and Therapy, 17th Edition, www.merck.com) and by consulting with medical professionals in the area of infant pain.

2. Procedures analysed under scope of current review.

Name	Description
Heel stick	Heel stick involves lancing of the lateral aspect of the infant's heel, squeezing the heel, and collecting the pooled blood.
Venipuncture (also known as venepuncture)	Puncture of a vein, especially for the withdrawal of blood or for intravenous medication.
Needlestick	The act of forcing a liquid (such as vaccines) into tissue using a sharp needle.
Diaper change (preterm only)	Care‐giving intervention in which soiled diapers are changed and causes distress that some suspect to be pain‐related.
Endotracheal suctioning	A component of bronchial hygiene for mechanical ventilation and involves the mechanical aspiration of pulmonary secretion from the intubated airway. Its primary purpose is to remove airway secretions in order to prevent obstructions.
Weighing procedure (preterm only)	Routine care procedure in which the preterm is weighed on a scale and causes distress that some suspect to be pain‐related.
Retinopathy of prematurity examination	Ophthalmological examination to detect retinopathy of prematurity (ROP), a potentially blinding eye disorder that primarily affects premature infants.
Peripheral cannulation	A cannula/catheter is inserted into a small peripheral vein for therapeutic purposes such as administration of medications, fluids and/or blood products.
Orogastric tube insertion	The placement of a tube into the stomach via the oropharynx to facilitate feeding in preterm infants.

The following painful procedures were used in the 138 included studies.

Heel stick (63 studies; Abbasoglu 2015; Akcan 2016; Alemdar 2017a; Angeles 2020; Axelin 2009; Aydin 2019; Badiee 2014; Bellieni 2001; Bellieni 2002; Bembich 2018; Blass 1999; Bo 2000; Campbell‐Yeo 2012; Campos 1994; Chang 2020; Chen 2017; Chen 2019; Chirico 2017; Corbo 2000; Corff 1995; Davari 2018; Dur 2020; Efendi 2018; Erkut 2017; Fearon 1997; Gao 2018; Gormally 2001; Greenberg 2002; Herrington 2007; Herrington 2014; Ho 2016; Im 2008; Jain 2006; Johnston 1997; Johnston 2007a; Johnston 2013; Kahraman 2020; KarabiyikOgurlu 2020; Kozub 2001; Leng 2016; Liaw 2010; Liaw 2012; Liaw 2013; Malakian 2017; McGinnis 2016; Mir 2018; Mirzarahimi 2013; Mitchell 2016; Morrow 2010; Ozkan 2019; Peng 2018; Perroteau 2018; Ranjbar 2020; Shu 2014; Stevens 1999; Sundaram 2013; Thakkar 2016; Ti 2015; Vu‐Ngoc 2020; Whipple 2004; Yilmaz 2010; Yilmaz 2020; Yu 2017).

Vaccine/vitamin needle (35 studies; Allen 1996; Basiri‐Moghadam 2014; Benjamin 2016; Bustos 2008; Caglar 2017; Chou 2012; Cohen 2002; Cohen 2006; Cramer‐Berness 2005a; Cramer‐Berness 2005b; Dabas 2019; Esfahani 2013; Göl 2017; Gray 2015; Harrington 2012; Hashemi 2016; Hillgrove Stuart 2013; Hogan 2014; Ipp 2004; Jose 2012; Karaca 2016; Khasanah 2019; Khosravan 2018; Koç 2015; Kristoffersen 2011; Kucukoglu 2015; Liaw 2011; Najafi 2019; Petrillo 2010; Pillai Riddell 2018; Shaw 1982; Taddio 2014; Taddio 2015; Taddio 2017; Taddio 2017a).

Venipuncture (20 studies; Alemdar 2017; Ayazi 2017; Bahorski 2015; Bauchner 1996; Baudesson 2017; Beiranvand 2020; Carbajal 1999; Carbajal 2003; Chik 2012; Chik 2017; Curtis 2007; Dehghani 2019; Dezhdar 2016; Elserafy 2009; Fitri 2020; Jebreili 2015; Liu 2010; Lopez 2015; Sadathosseini 2013; Secil 2014).

Diaper change (two studies; Comaru 2009; Sizun 2002).

Endotracheal suctioning technique (five studies; Alinejad‐Naeini 2014; Axelin 2006; Fatollahzade 2020; Taplak 2020; Ward‐Larson 2004).

Weighing procedure (two studies; Catelin 2005; Hill 2005).

Retinopathy of prematurityscreening procedure (three studies; Behnam Vashani 2015; Chuang 2019; Ren 2019).

Peripheral cannulation (two studies; Alemdar 2018; Alemdar 2020).

Orogastric tube insertion (one study; Apaydin Cirik 2020).

Unspecified blood sampling procedure (three studies; Azarmnejad 2015; Razaghi 2020; Roshanray 2020).

Unspecified invasive procedure (one study; Cetinkaya 2020).

Unspecified painful procedures (one study; Erdogan 2020).

Types of interventions

In total, we included 24 different types of interventions in this review. Of these, we meta‐analysed 18 types of interventions while six were only described qualitatively. We examined the efficacy of the different types of interventions separately for preterms, neonates, and older infants. Moreover, within each treatment x age analysis, we separated the data into pain phases (pain reactivity or immediate pain regulation). We did not include any of the 255 treatment arms from the 138 included studies twice within any single treatment x age x pain type analyses. 

Analysed interventions

1. Non‐nutritive sucking‐related (33 treatment arms total)

1. Preterm born neonate

   1. Pain reactivity: Bellieni 2001; Corbo 2000; Elserafy 2009; Elserafy 2009; Gao 2018; Kristoffersen 2011; Liaw 2010; Stevens 1999 

   2. Immediate pain regulation: Corbo 2000; Elserafy 2009; Elserafy 2009; Gao 2018; Liaw 2010; Liaw 2012; Whipple 2004

2. Full‐term born neonate

   1. Pain reactivity: Bellieni 2002; Carbajal 1999; Chang 2020; Liu 2010; Mirzarahimi 2013; Najafi 2019; Thakkar 2016; Yilmaz 2010

   2. Immediate pain regulation: Blass 1999; Bo 2000; Campos 1994; Im 2008; Liaw 2011; Liu 2010; Najafi 2019; Vu‐Ngoc 2020; Yilmaz 2010 

3. Full‐term born older infant

   1. Pain reactivity: none

   2. Immediate pain regulation: Curtis 2007

2. Swallowing water (four treatment arms total)

1. Preterm born neonate

   1. Pain reactivity: Elserafy 2009; Kristoffersen 2011

   2. Immediate pain regulation: Elserafy 2009

2. Full‐term born neonate: 

   1. Pain reactivity: none 

   2. Immediate pain regulation: none

3. Full‐term born older infant: 

   1. Pain reactivity: none

   2. Immediate pain regulation: Allen 1996

3. Rocking, holding or both (seven treatment arms total)

1. Preterm born neonate: 

   1. Pain reactivity: none

   2. Immediate pain regulation: none

2. Full‐term born neonate:

   1. Pain reactivity: Carbajal 2003; Gormally 2001; Roshanray 2020

   2. Immediate pain regulation: Campos 1994; Gormally 2001; Roshanray 2020

3. Full‐term born older infant: 

   1. Pain reactivity: Ipp 2004

   2. Immediate pain regulation: none

4. Swaddling (14 treatment arms total)

1. Preterm born neonate:

   1. Pain reactivity: Apaydin Cirik 2020; Dezhdar 2016; Ho 2016; Shu 2014; Stevens 1999

   2. Immediate pain regulation: Apaydin Cirik 2020; Dezhdar 2016; Fearon 1997; Ho 2016 

2. Full‐term born neonate:

   1. Pain reactivity: Erkut 2017; Hashemi 2016; Morrow 2010; Yilmaz 2020

   2. Immediate pain regulation: Erkut 2017

3. Full‐term born older infant: 

   1. Pain reactivity: none

   2. Immediate pain regulation: none

5. Facilitated tucking (25 treatment arms total)

1. Preterm born neonate:

   1. Pain reactivity: Alinejad‐Naeini 2014; Apaydin Cirik 2020; Axelin 2006; Axelin 2009; Comaru 2009; Comaru 2009; Davari 2018; Hill 2005; Lopez 2015; Ranjbar 2020; Sundaram 2013; Taplak 2020; Ward‐Larson 2004

   2. Immediate pain regulation: Apaydin Cirik 2020; Axelin 2006; Corff 1995; Davari 2018; Herrington 2007; Herrington 2014; Liaw 2012; Ranjbar 2020; Sundaram 2013; Taplak 2020

2. Full‐term born neonate:

   1. Pain reactivity: Kucukoglu 2015

   2. Immediate pain regulation: none

3. Full‐term born older infant:

   1. Pain reactivity: Khasanah 2019

   2. Immediate pain regulation: none

6. Touch/massage related (27 treatment arms total)

1. Preterm born neonate:

   1. Pain reactivity: Abbasoglu 2015; Chik 2012; Chik 2012; Fatollahzade 2020

   2. Immediate pain regulation: Dehghani 2019; Jain 2006

2. Full‐term born neonate:

   1. Pain reactivity: Chik 2017; Mir 2018; Mirzarahimi 2013; Ozkan 2019; Ozkan 2019; Roshanray 2020; Yu 2017

   2. Immediate pain regulation: Im 2008; Ozkan 2019; Ozkan 2019; Roshanray 2020

3. Full‐term born older infant:

   1. Pain reactivity: Beiranvand 2020; Esfahani 2013; Göl 2017; Karaca 2016, Khosravan 2018; Koç 2015 

   2. Immediate pain regulation: Beiranvand 2020; Göl 2017; Jose 2012; Karaca 2016

7. Toy distraction (eight treatment arms total)

1. Preterm born neonate:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

2. Full‐term born neonate:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

3. Full‐term born older infant:

   1. Pain reactivity: Basiri‐Moghadam 2014; Cramer‐Berness 2005a; Cramer‐Berness 2005b; Dabas 2019; Hillgrove Stuart 2013; Hillgrove Stuart 2013

   2. Immediate pain regulation: Hillgrove Stuart 2013; Hillgrove Stuart 2013

8. Structured parent involvement (14 treatment arms total)

1. Preterm born neonate: 

   1. Pain reactivity: none

   2. Immediate pain regulation: none

2. Full‐term born neonate:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

3. Full‐term born older infant:

   1. Pain reactivity: Bustos 2008; Cramer‐Berness 2005a; Cramer‐Berness 2005b; Pillai Riddell 2018; Pillai Riddell 2018; Taddio 2015; Taddio 2017; Taddio 2017a

   2. Immediate pain regulation: Bauchner 1996; Pillai Riddell 2018; Pillai Riddell 2018; Taddio 2015; Taddio 2017; Taddio 2017a

9. Sound reduction (three treatment arms total)

1. Preterm born neonate:

   1. Pain reactivity: Ayazi 2017; Kahraman 2020

   2. Immediate pain regulation: Ayazi 2017

2. Full‐term born neonate: 

   1. Pain reactivity: none

   2. Immediate pain regulation: none

3. Full‐term born older infant:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

10. Sound addition (16 treatment arms total)

1. Preterm born neonate:

   1. Pain reactivity: Alemdar 2017; Alemdar 2018; Chen 2019; Chirico 2017; Johnston 2007a; Kahraman 2020; Kahraman 2020; Taplak 2020

   2. Immediate pain regulation: Alemdar 2017; Alemdar 2018; Ren 2019; Taplak 2020

2. Full‐term born neonate:

   1. Pain reactivity: Azarmnejad 2015; Cetinkaya 2020

   2. Immediate pain regulation: none

3. Full‐term born older infant:

   1. Pain reactivity: Erdogan 2020

   2. Immediate pain regulation: Erdogan 2020

11. Smell addition (25 treatment arms total)

1. Preterm born neonate:

   1. Pain reactivity: Alemdar 2017a; Alemdar 2017a; Alemdar 2017a; Alemdar 2018; Alemdar 2020; Jebreili 2015; Jebreili 2015; Taplak 2020 

   2. Immediate pain regulation: Alemdar 2017a; Alemdar 2017a; Alemdar 2017a; Alemdar 2018; Alemdar 2020; Jebreili 2015; Jebreili 2015; Taplak 2020

2. Full‐term born neonate:

   1. Pain reactivity: Akcan 2016; Akcan 2016; Akcan 2016; Razaghi 2020

   2. Immediate pain regulation: Akcan 2016; Akcan 2016; Akcan 2016; Sadathosseini 2013; Sadathosseini 2013

3. Full‐term born older infant:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

12. Light reduction (four treatment arms total)

1. Preterm born neonate:

   1. Pain reactivity: Alemdar 2017; Alemdar 2018

   2. Immediate pain regulation: Alemdar 2017; Alemdar 2018

2. Full‐term born neonate:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

3. Full‐term born older infant:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

13. Heat addition (four treatment arms total)

1. Preterm born neonate: 

   1. Pain reactivity: Shu 2014

   2. Immediate pain regulation: none

2. Full‐term born neonate:

   1. Pain reactivity: Aydin 2019; KarabiyikOgurlu 2020; Mir 2018

   2. Immediate pain regulation: none

3. Full‐term born older infant:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

14. Cold addition (three treatment arms total)

1. Preterm born neonate: 

   1. Pain reactivity: none

   2. Immediate pain regulation: none

2. Full‐term born neonate:

   1. Pain reactivity: Malakian 2017; Najafi 2019

   2. Immediate pain regulation: Najafi 2019

3. Full‐term born older infant:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

15. Multisensory bundle (13 treatment arms total)

1. Preterm born neonate:

   1. Pain reactivity: Behnam Vashani 2015; Catelin 2005; Chuang 2019; Johnston 1997; Sizun 2002

   2. Immediate pain regulation: Behnam Vashani 2015; Catelin 2005; Chuang 2019; Efendi 2018

2. Full‐term born neonate:

   1. Pain reactivity: Bellieni 2002; Chou 2012

   2. Immediate pain regulation: none

3. Full‐term born older infant:

   1. Pain reactivity: Khosravan 2018

   2. Immediate pain regulation: Harrington 2012

16. Additive non‐pharmacological intervention on sweet solution (27 treatment arms total)

1. Preterm born neonate

   1. Pain reactivity: Angeles 2020; Campbell‐Yeo 2012; Chen 2017; Fitri 2020; Gao 2018; Kristoffersen 2011; Liaw 2013; Petrillo 2010; Ti 2015

   2. Immediate pain regulation: Campbell‐Yeo 2012; Gao 2018; Liaw 2013; Ren 2019

2. Full‐term born neonate:

   1. Pain reactivity: Bembich 2018; McGinnis 2016; Mitchell 2016; Petrillo 2010; Thakkar 2016

   2. Immediate pain regulation: Gray 2015; Greenberg 2002; Leng 2016; Leng 2016; McGinnis 2016; Mitchell 2016

3. Full‐term born older infant

   1. Pain reactivity: Hogan 2014; Taddio 2014

   2. Immediate pain regulation: Harrington 2012

17. Additive non‐pharmacological intervention on non‐nutritive sucking (12 treatment arms total)

1. Preterm born neonate:

   1. Pain reactivity: Baudesson 2017; Dur 2020; Dur 2020; Liaw 2013; Peng 2018; Perroteau 2018

   2. Immediate pain regulation: Dur 2020; Dur 2020; Liaw 2013; Peng 2018

2. Full‐term born neonate:

   1. Pain reactivity: Mitchell 2016

   2. Immediate pain regulation: Mitchell 2016

3. Full‐term born older infant:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

18. Additive non‐pharmacological intervention on swaddling (three treatment arms total)

1. Preterm born neonate: 

   1. Pain reactivity: none

   2. Immediate pain regulation: none

2. Full‐term born neonate:

   1. Pain reactivity: Caglar 2017; Yilmaz 2020

   2. Immediate pain regulation: Caglar 2017

3. Full‐term born older infant: 

   1. Pain reactivity: none

   2. Immediate pain regulation: none

Interventions included and described qualitatively

1. Therapeutic touch (three treatment arms total)

1. Preterm born neonate:

   1. Pain reactivity: Johnston 2013

   2. Immediate pain regulation: Johnston 2013 

2. Full‐term born neonate:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

3. Full‐term born older infant:

   1. Pain reactivity: Kozub 2001

   2. Immediate pain regulation: none

2. Vibration therapy (four treatment arms total)

1. Preterm born neonate:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

2. Full‐term born neonate:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

3. Full‐term born older infant:

   1. Pain reactivity: Benjamin 2016; Secil 2014

   2. Immediate pain regulation: Benjamin 2016; Secil 2014

3. Video distraction (two treatment arms total)

1. Preterm born neonate:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

2. Full‐term born neonate:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

3. Full‐term born older infant

   1. Pain reactivity: Cohen 2002

   2. Immediate pain regulation: Cohen 2006

4. Parent presence (two treatment arms total)

1. Preterm born neonate: 

   1. Pain reactivity: none

   2. Immediate pain regulation: none

2. Full‐term born neonate:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

3. Full‐term born older infant

   1. Pain reactivity: Shaw 1982

   2. Immediate pain regulation: Bauchner 1996

5. Co‐bedding (one treatment arm total)

1. Preterm born neonate: 

   1. Pain reactivity: Badiee 2014

   2. Immediate pain regulation: none

2. Full‐term born neonate:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

3. Full‐term born older infant:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

6. Additive non‐pharmacological intervention on topical anaesthetic (one treatment arm total)

1. Preterm born neonate: 

   1. Pain reactivity: none

   2. Immediate pain regulation: none

2. Full‐term born neonate:

   1. Pain reactivity: none

   2. Immediate pain regulation: none

3. Full‐term born older infant: 

   1. Pain reactivity: none

   2. Immediate pain regulation: Bahorski 2015

Treatment settings

The treatment interventions described in the 138 included studies occurred in the following settings.

Inpatient hospital (113 studies; Abbasoglu 2015; Akcan 2016; Alemdar 2017; Alemdar 2017a; Alemdar 2018; Alemdar 2020; Alinejad‐Naeini 2014; Allen 1996; Angeles 2020; Apaydin Cirik 2020; Axelin 2006; Axelin 2009; Ayazi 2017; Aydin 2019; Azarmnejad 2015; Badiee 2014; Bahorski 2015; Bauchner 1996; Baudesson 2017; Behnam Vashani 2015; Beiranvand 2020; Bellieni 2001; Bellieni 2002; Bembich 2018; Blass 1999; Bo 2000; Caglar 2017; Campbell‐Yeo 2012; Campos 1994; Carbajal 1999; Carbajal 2003; Catelin 2005; Cetinkaya 2020; Chang 2020; Chen 2017; Chen 2019; Chik 2012; Chik 2017; Chirico 2017; Chou 2012; Chuang 2019; Comaru 2009; Corbo 2000; Corff 1995; Curtis 2007; Davari 2018; Dehghani 2019; Dezhdar 2016; Dur 2020; Efendi 2018; Elserafy 2009; Erdogan 2020; Erkut 2017; Fatollahzade 2020; Fearon 1997; Fitri 2020; Gao 2018; Gormally 2001; Gray 2015; Greenberg 2002; Harrington 2012; Hashemi 2016; Herrington 2007; Herrington 2014; Hill 2005; Ho 2016; Im 2008; Jain 2006; Jebreili 2015; Johnston 1997; Johnston 2007a; Johnston 2013; Kahraman 2020; KarabiyikOgurlu 2020; Kozub 2001; Kristoffersen 2011; Kucukoglu 2015; Leng 2016; Liaw 2010; Liaw 2011; Liaw 2012; Liaw 2013; Liu 2010; Lopez 2015; Malakian 2017; McGinnis 2016; Mirzarahimi 2013; Mitchell 2016; Morrow 2010; Najafi 2019; Ozkan 2019; Peng 2018; Perroteau 2018; Ranjbar 2020; Razaghi 2020; Ren 2019; Roshanray 2020; Sadathosseini 2013; Secil 2014; Shu 2014; Sizun 2002; Stevens 1999; Sundaram 2013; Taddio 2015; Taplak 2020; Thakkar 2016; Ti 2015; Vu‐Ngoc 2020; Ward‐Larson 2004; Whipple 2004; Yilmaz 2010; Yilmaz 2020; Yu 2017).

Outpatient medical clinic (25 studies; Basiri‐Moghadam 2014; Benjamin 2016; Bustos 2008; Cohen 2002; Cohen 2006; Cramer‐Berness 2005a; Cramer‐Berness 2005b; Dabas 2019; Esfahani 2013; Göl 2017; Hillgrove Stuart 2013; Hogan 2014; Ipp 2004; Jose 2012; Karaca 2016; Khasanah 2019; Khosravan 2018; Koç 2015; Mir 2018; Petrillo 2010; Pillai Riddell 2018; Shaw 1982; Taddio 2014; Taddio 2017; Taddio 2017a).

Pain measurement

The pain measurements used in the 138 included studies were as follows.

Premature Infant Pain Profile (PIPP) (50 studies: Abbasoglu 2015; Alemdar 2017a; Alemdar 2018; Alemdar 2020; Alinejad‐Naeini 2014; Angeles 2020; Apaydin Cirik 2020; Axelin 2009; Ayazi 2017; Badiee 2014; Baudesson 2017; Behnam Vashani 2015; Bellieni 2001; Campbell‐Yeo 2012; Chen 2017; Chen 2019; Chik 2012; Chik 2017; Chirico 2017; Chuang 2019; Davari 2018; Dezhdar 2016; Efendi 2018; Elserafy 2009; Fatollahzade 2020; Fitri 2020; Gao 2018; Hill 2005; Ho 2016; Jebreili 2015; Johnston 2007a; Johnston 2013; Kozub 2001; Kristoffersen 2011; Liaw 2010; Liaw 2012; Lopez 2015; Mirzarahimi 2013; Mitchell 2016; Peng 2018; Perroteau 2018; Petrillo 2010; Ranjbar 2020; Ren 2019; Sizun 2002; Sundaram 2013; Stevens 1999; Taplak 2020; Thakkar 2016; Ward‐Larson 2004).

Neonatal Infant Pain Scale (NIPS) (31 studies: Akcan 2016; Alemdar 2017; Axelin 2006; Aydin 2019; Azarmnejad 2015; Bembich 2018; Bo 2000; Caglar 2017; Catelin 2005; Cetinkaya 2020; Dehghani 2019; Dur 2020; Erkut 2017; Esfahani 2013; Göl 2017; Im 2008; Jain 2006; Kahraman 2020; KarabiyikOgurlu 2020; Karaca 2016; Kucukoglu 2015; Liu 2010; Mir 2018; Morrow 2010; Najafi 2019; Ozkan 2019; Roshanray 2020; Shu 2014; Yilmaz 2010; Yilmaz 2020; Yu 2017).

Duration of cry (nine studies: Allen 1996; Blass 1999; Campos 1994; Corff 1995; Gray 2015; Greenberg 2002; Herrington 2007; Herrington 2014; Sadathosseini 2013).

Neonatal Facial Coding System (NFCS) (10 studies: Bustos 2008; Chou 2012; Comaru 2009; Fearon 1997; Gormally 2001; Hashemi 2016; Ipp 2004; Johnston 1997; Leng 2016; Liaw 2011).

Modified Behavioural Pain Scale (MBPS) (12 studies: Beiranvand 2020; Cohen 2002; Cramer‐Berness 2005a; Cramer‐Berness 2005b; Hillgrove Stuart 2013; Hogan 2014; Khosravan 2018; Pillai Riddell 2018; Taddio 2014; Taddio 2015; Taddio 2017; Taddio 2017a).

Douleur Aiguë Nouveau‐né (DAN) (four studies: Bellieni 2002; Carbajal 1999; Carbajal 2003; Razaghi 2020).

N‐PASS (three studies: Chang 2020; McGinnis 2016; Vu‐Ngoc 2020).

Active/sleep state (one study: Liaw 2013).

Behavioural Observation Pain Scale (one study: Jose 2012).

Brazelton Neonatal Behavioral Assessment Scale (BNBAS) (one study: Corbo 2000).

Face Legs Arms Cry Consolability Scale (FLACC) (seven studies: Benjamin 2016; Curtis 2007; Dabas 2019; Erdogan 2020; Khasanah 2019; Koç 2015; Secil 2014).

Infant behavioural state (one study: Whipple 2004).

Infant's Pain Questionnaire (one study: Basiri‐Moghadam 2014).

Cry frequency (one study: Bauchner 1996).

Children's Hospital of Eastern Ontario Pain Scale (CHEOPS) (one study; Bahorski 2015).

CRIES (one study; Malakian 2017).

Modified Frankl Scale (one study; Shaw 1982).

Modified Riley Infant Pain Scoring (one study; Harrington 2012).

Measure of Adult and Infant Soothing and Distress(MAISD) (one study; Cohen 2006).

Grimace behaviours (one study; Ti 2015).

Excluded studies

In total, there are 305 excluded studies.

For this update, 11 authors (OB, IS, CC, HG, SB, MD, NR, SAK, KT, DL, LU) screened 11,275 abstracts based on the search strategy. From these 11,275 abstracts, we selected 417 for further review. Three authors (RPR, OB, IS) conducted a full‐text review of these 417 articles, leading to the exclusion of 305 studies because of wrong intervention (N = 127), wrong age/age could not be separated (N = 71), no acute painful procedures (N = 24), no control group/active control group (N = 35), no behavioural/multidimensional pain measure (N = 11), no pain outcome/pain not measured during procedure (N = 15), no RCT (N = 13), wrong study type (N = 9) (see Characteristics of excluded studies table and Characteristics of studies awaiting classification (N = 36)). We placed studies in awaiting classification due to not finding a translator for non‐English articles (N = 2), not finding the full‐text article (N = 1), and eligible studies from the October 2020 to July 2022 updated search awaiting assessment in future updates of the review (N = 33). 

Risk of bias in included studies

As described earlier, we calculated risk of bias using domains from Chapter 8 in the 2011 Cochrane Handbook for Systematic Reviews of Interventions (see Figure 2 and Figure 3) (Higgins 2011a; Sterne 2019). The risk of bias breakdown for all 138 included studies is provided below. Amongst the domains, the largest area of challenge (over 50%) to bias related to descriptions of blinding procedures. Adherence to CONSORT reporting procedures by study authors continues to be a challenge in this updated version of the review.

2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

3.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Overall risk of bias

We rated 103 studies as having a high risk of bias, 30 studies as having an unclear risk of bias, and five studies as having a low risk of bias. Further details are provided below. Figure 2 and Figure 3 summarise the ratings for each of the included studies.

High risk of bias (103 studies; Alemdar 2017; Alemdar 2017a; Alemdar 2018; Alemdar 2020; Alinejad‐Naeini 2014; Allen 1996; Angeles 2020; Apaydin Cirik 2020; Axelin 2006; Ayazi 2017; Aydin 2019; Azarmnejad 2015; Badiee 2014; Bahorski 2015; Bauchner 1996; Behnam Vashani 2015; Benjamin 2016; Beiranvand 2020; Bellieni 2001; Bembich 2018; Blass 1999; Bo 2000; Bustos 2008; Caglar 2017; Campbell‐Yeo 2012; Carbajal 1999; Carbajal 2003; Cetinkaya 2020; Chang 2020; Chen 2019; Chik 2017; Chirico 2017; Chou 2012; Chuang 2019; Cohen 2002; Cohen 2006; Corbo 2000; Corff 1995; Cramer‐Berness 2005a; Dabas 2019; Davari 2018; Dezhdar 2016; Dur 2020; Efendi 2018; Erdogan 2020; Erkut 2017; Esfahani 2013; Fatollahzade 2020; Fitri 2020; Gao 2018; Göl 2017; Gormally 2001; Gray 2015; Greenberg 2002; Harrington 2012; Hashemi 2016; Hillgrove Stuart 2013; Ho 2016; Hogan 2014; Im 2008; Ipp 2004; Johnston 1997; Johnston 2007a; Jose 2012; Kahraman 2020; KarabiyikOgurlu 2020; Karaca 2016; Khosravan 2018; Koç 2015; Kristoffersen 2011; Kucukoglu 2015; Leng 2016; Liaw 2010; Liaw 2011; Liaw 2012; Liaw 2013; Liu 2010; Lopez 2015; Malakian 2017; McGinnis 2016; Mir 2018; Morrow 2010; Ozkan 2019; Peng 2018; Perroteau 2018; Petrillo 2010; Ranjbar 2020; Razaghi 2020; Roshanray 2020; Secil 2014; Shaw 1982; Sizun 2002; Stevens 1999; Taddio 2015; Taplak 2020; Thakkar 2016; Ti 2015; Vu‐Ngoc 2020; Ward‐Larson 2004; Whipple 2004; Yilmaz 2010; Yilmaz 2020; Yu 2017).

Unclear risk of bias (30 studies; Abbasoglu 2015; Akcan 2016; Axelin 2009; Basiri‐Moghadam 2014; Baudesson 2017; Bellieni 2002; Campos 1994; Catelin 2005; Chik 2012; Comaru 2009; Cramer‐Berness 2005b; Curtis 2007; Dehghani 2019; Elserafy 2009; Fearon 1997; Herrington 2007; Herrington 2014; Hill 2005; Jain 2006; Jebreili 2015; Johnston 2013; Khasanah 2019; Kozub 2001; Mirzarahimi 2013; Mitchell 2016; Najafi 2019; Ren 2019; Sadathosseini 2013; Shu 2014; Sundaram 2013).

Low risk of bias (five studies; Chen 2017; Pillai Riddell 2018; Taddio 2014; Taddio 2017; Taddio 2017a).

Allocation

Random sequence generation

All 138 studies were randomised. However, 23 studies did not adequately describe the method used to generate the random sequence and so we identified them as having unclear risk. Furthermore, we judged eight studies to be at high risk for this domain due to inadequate generation of a randomised sequence (e.g. randomisation based on week of admission). We deemed the remaining 107 studies to be of low risk for this domain as they described an adequate method of sequence generation.

Allocation concealment

Sixty‐seven studies did not adequately describe how the allocation of the sequence was concealed and we judged them to be at unclear risk of bias for this domain. We deemed 23 studies to be at high risk due to inadequate concealment of allocation prior to assignment. The remaining 48 studies described an adequate method of allocation concealment prior to assignment and we thus deemed them to be at low risk for this domain.

Blinding

Performance bias

Thirty‐six studies did not adequately describe the method by which personnel were blinded during the study and we judged them to be at unclear risk of bias for this domain. We deemed 76 studies to be at high risk due to no blinding of personnel, which could have introduced performance bias. The remaining 26 studies described an adequate method of personnel blinding during the study and we thus deemed them to be at low risk for this domain. 

Detection bias

Forty‐two studies did not adequately describe how the outcome assessors were blinded and we judged them to be at an unclear risk of bias. We deemed 58 studies at high risk due to inadequate methods of blinding of outcome assessors. The remaining 38 studies described an adequate method through which outcome assessors were blinded and we judged them to be at a low risk for this domain.

Incomplete outcome data

Eighteen studies did not adequately describe the amount, nature, or handling of incomplete outcome data, and we judged them to be at unclear risk of bias for this domain. We deemed 14 studies to be at a high risk of bias due to an inadequate method of handling incomplete outcome data. The remaining 106 studies described an adequate method of handling incomplete outcome data and we judged them to be at a low risk of bias for this domain.

Selective reporting

Ninety‐nine studies did not provide their pre‐specified outcomes of interest, and so we were unable to judge whether selective outcome reporting had occurred, leading to us judging these studies at an unclear risk of bias for this domain. The remaining 39 studies made their study protocols available, and we accessed the study protocol and found that all pre‐specified outcomes were reported in the study, leading to judging these studies to be at a low risk of bias for this domain. No studies were identified as having high risk of bias for this domain.

Other potential sources of bias

We deemed four studies to be at an unclear risk of bias due to suspected methodological problems not covered above (i.e. potential issues with reliability of pain behaviours coding, exclusion criteria that could have introduced bias), but insufficient evidence was provided to clearly make a 'low' or 'high' risk of bias judgement. We deemed six studies to be at a high risk of bias for this domain as they had at least one other important source of bias (e.g. their execution of a methodological confound, such as not keeping infant position consistent across groups, which challenged our ability to interpret the true impact of the treatment). The remaining 128 studies appeared to be free of other sources of bias, and we judged them to be at a low risk for this domain.

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4; Table 5

As aforementioned, to examine the efficacy of each intervention (see Types of interventions above), we organised the results by pain management interventions (24 distinct interventions), age group (preterm born neonate, full‐term born neonate, full‐term born older infant), and pain phase (pain reactivity, immediate pain regulation). We reported all 138 included studies in this format below. Of these 138 studies, we meta‐analysed 115 and described 23 in a qualitative manner. 

Of all 24 included interventions, we meta‐analysed 18 interventions (44 separate analyses), which are presented first below, followed by the six included interventions that we described qualitatively. We referred to studies included in the meta‐analysis as 'analysed' studies. Studies included but not meta‐analysed were referred to as 'included qualitatively'.

Ultimately, no analyses were based on sufficient evidence. We based sufficient evidence for an intervention on the following criteria: the analysis must be based on a total treatment number ('n') that exceeds 30x where x is the number of studies, the majority of studies must have low risk of bias overall, and evidence must come from at least two separate research groups. We based the criteria on recommendations from Cochrane, GRADE criteria, and the American Psychological Association (Chambless 2001). Please refer to the 'Summary of overall quantitative findings' table for a summary (Table 6).

For our secondary outcome (i.e. adverse events), 62 studies did not report on these. Seventy studies explicitly reported that no adverse events occurred. Six studies reported adverse events (vomiting, abdominal distension, transient skin reactions, desaturations, infants becoming distressed, septicaemia). Adverse events were reported in both the control and treatment groups (Axelin 2006; Curtis 2007; Gao 2018; Taddio 2014; Taddio 2017; Thakkar 2016).

1. Non‐nutritive sucking versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Seven studies (eight treatment arms) analysed the effects of non‐nutritive sucking compared to no‐treatment control on preterm pain reactivity (see Analysis 1.1). The total number of participants in the analysis was 597 (treatment = 299, control = 298). The mean sample size across studies was 53 (standard deviation (SD) = 42; range = 17 to 122). Most studies were cross‐over trials (six out of eight treatment arms). One study contributed two treatment arms to the analysis: a pacifier treatment arm and a pacifier plus water treatment arm (Elserafy 2009). Non‐nutritive sucking for preterm pain reactivity may reduce pain reactivity based on evidence of very low certainty (standardised mean difference (SMD) ‐0.57, 95% confidence interval (CI) ‐1.03 to ‐0.11; I² = 93%), which we downgraded twice for very serious study limitations as most studies had a high risk of bias overall, and once for inconsistency (see Table 1). 

1.1. Analysis.

Comparison 1: Non‐nutritive sucking X preterm neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Six studies analysed the effects of sucking on preterm immediate pain regulation, with one study contributing two arms (Elserafy 2009) (see Analysis 1.2). The total number of participants in the analysis was 379 (treatment = 190, control = 189). The mean sample size across studies was 47 (SD = 28; range = 26 to 104). Most studies were cross‐over trials (four out of seven treatment arms). There is evidence that non‐nutritive sucking may improve immediate pain regulation compared to no‐treatment control in preterm infants (SMD ‐0.61, 95% CI ‐0.95 to ‐0.27; I² = 81%), based on evidence of very low certainty, which we downgraded twice for very serious study limitations as most studies included had a high risk of bias overall, and once for inconsistency (see Table 1).

1.2. Analysis.

Comparison 1: Non‐nutritive sucking X preterm neonate, Outcome 2: Immediate pain regulation

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Eight studies analysed the effect of sucking on pain reactivity for full‐term neonates. The SMD is ‐1.13 (95% CI ‐1.57 to ‐0.68; I² = 82%), suggesting that for full‐term neonates, sucking may reduce pain reactivity compared to no‐treatment control (see Analysis 2.1). The total number of analysed participants was 545 (treatment = 271, control = 274). The mean sample size across studies was 68 (SD = 20; range = 40 to 101). All studies were between‐group analyses. One study involved stimulating sucking with water for 30 seconds after the needle, whereas the others involved placing a pacifier in the baby’s mouth. This was based on evidence of very low certainty, which we downgraded twice for very serious study limitations due to the overall risk of bias and once for inconsistency (see Table 2). 

2.1. Analysis.

Comparison 2: Non‐nutritive sucking X full‐term neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Nine studies analysed the effect of sucking on the immediate regulation of pain for full‐term neonates (see Analysis 2.2). An overall SMD of ‐1.49 (95% CI ‐2.20 to ‐0.78; I² = 92%) suggests that there is evidence of very low certainty that sucking may be efficacious for improving immediate pain regulation in full‐term neonates (see Table 2). We downgraded the certainty of evidence twice for very serious study limitations (overall risk of bias) and once for inconsistency. The total number of participants in the analyses was 536 (treatment = 267, control = 269). The mean sample size across studies was 58 (SD = 27; range = 20 to 110). Most studies were between‐group analyses (eight out of nine treatment arms).

2.2. Analysis.

Comparison 2: Non‐nutritive sucking X full‐term neonate, Outcome 2: Immediate pain regulation

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

One study with 41 participants (treatment = 22, control = 19), included qualitatively, reported no evidence of a difference in efficacy between non‐nutritive sucking and no‐treatment at reducing older infants' behavioural pain scores (Curtis 2007). It was proposed that the efficacy of pacifiers diminishes with increasing age beyond three months old. We downgraded the certainty of evidence by two levels for imprecision (once since only one study reported on this outcome and once because the treatment n of this study did not exceed the threshold for imprecision of 30x).

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

Seven studies reported on adverse events. While six studies did not report any adverse events, Gao 2018 described that a participant in the non‐nutritive sucking intervention group vomited (see Table 1). This was based on very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias) and once for inconsistency.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Ten studies reported on adverse events. Thakkar 2016 reported that one participant in the treatment group and two participants in the control group desaturated during the study. The remaining studies did not report any adverse events (see Table 2). This was based on very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias) and once for imprecision.

Full‐term born older infants (over one month to 36 months of age)

Curtis 2007 reported that three participants vomited following the intervention. We deemed the certainty of this evidence to be low as it was downgraded twice for imprecision.

Summary of treatment effects: non‐nutritive sucking

Preterm born neonates

The results of this meta‐analysis suggest that there is very uncertain evidence that sucking may be efficacious in reducing both pain reactivity and improving immediate pain regulation in preterm neonates. An analysis of studies that found evidence for the efficacy of non‐nutritive sucking over no‐treatment suggests that pain relief will be maximised if sucking begins at least three minutes prior to the painful stimuli. One study reported an adverse event (vomiting) following this intervention in preterm neonates (Gao 2018). 

Full‐term born neonates

The results suggest that sucking may reduce pain reactivity and improve immediate pain regulation in full‐term born neonates, based on evidence of very low certainty. One study reported an adverse event (desaturation) following this intervention in full‐term neonates (Thakkar 2016). 

Full‐term born older infants

One included study described qualitatively did not find evidence of a difference between non‐nutritive sucking and no‐treatment, based on low‐certainty evidence (Curtis 2007). This study reported adverse events (vomiting) in three participants following this intervention. 

2. Swallowing water versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Two studies analysed the effect of swallowing water on pain reactivity in preterm neonates (see Analysis 3.1). The total number of participants was 120 (control = 60, treatment = 60). The mean sample size across studies was 30 (SD = 8; range = 24 to 36). Both treatment arms were cross‐over studies. These studies found that water was not more efficacious than a no‐treatment control in reducing preterm pain reactivity (SMD 1.10, 95% CI ‐1.56 to 3.75; I² = 98%), based on evidence of low certainty, which we downgraded once for serious study limitations (overall risk of bias) and once for inconsistency.

3.1. Analysis.

Comparison 3: Swallowing water X preterm neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

One included cross‐over trial, described qualitatively (36 participants; control = 36, treatment = 36) (Elserafy 2009), was the only study with a treatment arm examining the effects of swallowing water on improving pain regulation in this age group, concluding that swallowing water is not efficacious at improving pain regulation based on evidence of moderate certainty, which we downgraded once for imprecision.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus

One study with 64 participants (treatment = 32, control = 32), included qualitatively, reported that administering water prior to a needle improved immediate pain regulation in older infants only when one needle is administered (Allen 1996). This was based on very low‐certainty evidence, downgraded twice for very serious study limitations (overall risk of bias) and once for imprecision. This study also contained neonates but, as the sample was predominantly older infants at or above two months of age, we categorised the age group as full‐term born older infants.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

Only one study explicitly mentioned adverse effects, and it did not find any during the study (36 participants; Elserafy 2009). We deemed the certainty of this evidence to be moderate, downgraded once for imprecision.

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Allen 1996 (64 participants) did not report any adverse effects, based on very low‐certainty evidence due to two downgrades for very serious study limitations (overall risk of bias) and one for imprecision.

Summary of treatment effects: swallowing water

Preterm born neonates

There is evidence of low to moderate certainty showing that water is an ineffective intervention for pain reactivity and immediate pain regulation in preterm neonates. One study reported that no adverse events occurred. 

Full‐term born neonates

No studies reported on this outcome.

Full‐term born older infants 

Allen 1996 suggested that swallowing water may be efficacious at improving pain regulation when only one needle is administered. This study reported that no adverse events occurred. 

The above studies used water as a treatment arm (in comparison to a ‘no‐treatment’ control), while most other studies in the literature used water as the ‘no‐treatment’ control group. Given the more common use of water in the literature and the limited evidence in every age group of its inefficacy, it is not recommended that further research explores swallowing water as a treatment for young child procedural pain.

3. Rocking/holding versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Three studies analysed the effect of rocking/holding on the pain reactivity of full‐term neonates following a painful procedure (see Analysis 4.1). The total number of participants was 221 (control = 111, treatment = 110). The mean sample size across studies was 74 (SD = 28; range = 41 to 90). All treatment arms were between‐group. The pooled effects found that rocking/holding alone was not more efficacious than a no‐treatment control group in reducing pain reactivity in full‐term neonates (SMD ‐0.09, 95% CI ‐0.61 to 0.43; I² = 72%), based on evidence of very low certainty, which we downgraded twice for very serious study limitations (overall risk of bias) and once for inconsistency.

4.1. Analysis.

Comparison 4: Rocking/holding X full‐term neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Three studies analysed the effect of rocking or holding on immediate pain regulation of full‐term neonates (see Analysis 4.2). The total number of participants was 171 (control = 86, treatment = 85). The mean sample size across studies was 57 (SD = 29; range = 40 to 90). All treatment arms were between‐group. The overall pooled effect showed that rocking/holding may be more efficacious than a no‐treatment control at improving immediate pain regulation for full‐term neonates (SMD ‐0.84, 95% CI ‐1.15 to ‐0.53; I² = 0%), based on evidence of very low certainty, which we downgraded twice for very serious study limitations (overall risk of bias) and once for imprecision.

4.2. Analysis.

Comparison 4: Rocking/holding X full‐term neonate, Outcome 2: Immediate pain regulation

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

One study with 106 participants (treatment = 56, control = 50), included qualitatively, concluded that holding is not more efficacious than standard care (in this case supine positioning) in reducing pain reactivity in healthy infants aged two to six months (Ipp 2004). This was based on low‐certainty evidence, downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus

No studies reported on this outcome.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Two studies reported on this outcome and neither reported any adverse effects, based on moderate‐certainty evidence due to one downgrade for serious study limitations (overall risk of bias).

Full‐term born older infants (over one month to 36 months of age)

Ipp 2004 did not report any adverse effects. This was based on low‐certainty evidence, downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Summary of treatment effects: rocking/holding

Preterm born neonates

No studies reported on this outcome.

Full‐term born neonates

The results indicated very low‐certainty evidence to suggest that rocking/holding was not efficacious for pain reactivity in full‐term neonates. Based on very uncertain evidence, rocking/holding may improve immediate pain regulation for full‐term neonates. No studies reported adverse events. 

Full‐term born older infants

An included study described qualitatively concluded that holding is not an efficacious analgesic in older infants (Ipp 2004). This study did not report any adverse events.

4. Swaddling versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Five studies analysed the effect of swaddling on pain reactivity in preterm neonates (see Analysis 5.1). Swaddling was found not efficacious in reducing pain reactivity in preterm neonates (SMD ‐0.60, 95% CI ‐1.23 to 0.04; I² = 91%). The total number of analysed participants was 468 (control = 236, treatment = 232). The mean sample size across studies was 69 (SD = 30; range = 50 to 122). Four studies were between‐group studies and one was a cross‐over. We deemed this to be very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias) and once for inconsistency (see Table 3). It should be noted that of the two studies that did not find an effect of swaddling on pain reactivity, the sample in Stevens 1999, unlike the other studies, consisted of very low birth weight (VLBW) infants in the neonatal intensive care unit (NICU), a particularly vulnerable population.

5.1. Analysis.

Comparison 5: Swaddling X preterm neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Four studies analysed whether swaddling was efficacious for immediate pain regulation of preterm neonates (see Analysis 5.2). The total number of analysed participants was 204 (control = 104, treatment = 100). The mean sample size across studies was 47 (SD = 22; range = 15 to 57). Three studies were between‐group and one was a cross‐over study. Swaddling may improve immediate pain regulation in preterm neonates (SMD ‐1.21, 95% CI ‐2.05 to ‐0.38; I² = 89%), based on evidence of very low certainty downgraded once for serious study limitations (overall risk of bias), once for inconsistency, and once for imprecision (see Table 3).

5.2. Analysis.

Comparison 5: Swaddling X preterm neonate, Outcome 2: Immediate pain regulation

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Four studies analysed whether swaddling was efficacious on pain reactivity in full‐term neonates (see Analysis 6.1). The total number of participants was 263 (control = 130, treatment = 133). The mean sample size across studies was 66 (SD = 17; range = 42 to 80). All four studies were between‐group. Swaddling may reduce pain reactivity in full‐term neonates (SMD ‐0.89, 95% CI ‐1.24 to ‐0.55; I² = 44%), based on low‐certainty evidence, which we downgraded twice for very serious study limitations due to overall risk of bias.

6.1. Analysis.

Comparison 6: Swaddling X full‐term neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

One study with 74 participants (treatment = 37, control = 37), included qualitatively, concluded that swaddling may be efficacious at improving pain regulation in full‐term neonates during heel lances (Erkut 2017). This was based on low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

The five studies reporting on this outcome did not observe any adverse events. We deemed this to be low‐certainty evidence due to two downgrades for very serious study limitations for overall risk of bias (see Table 3).

Full‐term born neonates (infants born at 37 weeks until one month of age)

Only one study reported on this outcome, and it did not find any adverse events. We judged the evidence to be very low certainty due to one downgrade for serious study limitations (overall risk of bias) and two downgrades for imprecision.

Full‐term born older infants (over one month to 36 months of age)

No studies reported on this outcome.

Summary of treatment effects: swaddling

Preterm born neonates

Evidence of very low certainty suggests that swaddling is not an efficacious intervention for pain reactivity in preterm neonates, whereas the efficacy of swaddling in improving immediate pain regulation in preterm neonates is supported by very low‐certainty evidence. No studies reported adverse events. 

Full‐term born neonates

Swaddling may reduce pain reactivity in full‐term neonates based on low‐certainty evidence. One study included and described qualitatively showed that swaddling may improve immediate pain regulation, which we also deemed low‐certainty evidence. No studies reported adverse events. 

Full‐term born older infants

No studies reported on this outcome.

5. Facilitated tucking versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Twelve studies (13 treatment arms) analysed the effect of facilitated tucking on pain reactivity in preterm neonates (see Analysis 7.1). One study reported two groups separately, which both received tucking and nesting with separate control groups (Comaru 2009). Thus, we entered these two groups into the analysis as separate studies since they used different participants. The total number of analysed participants was 733 (control = 367, treatment = 366). The mean sample size across studies was 34 (SD = 16; range = 12 to 65). Out of the 12 studies, nine were cross‐over. Facilitated tucking may be efficacious at reducing pain reactivity in preterm neonates (SMD ‐1.01, 95% CI ‐1.44 to ‐0.58; I² = 93%), based on very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias) and once for inconsistency (see Table 4).

7.1. Analysis.

Comparison 7: Facilitated tucking X preterm neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Ten studies analysed whether facilitated tucking was efficacious for immediate pain regulation of preterm neonates (see Analysis 7.2). The total number of analysed participants was 557 (control = 279, treatment = 278). The mean sample size across studies was 33 (SD = 19; range = 11 to 65). Out of 10 studies, eight were cross‐over design. Facilitated tucking may be efficacious for improving immediate pain regulation in preterm neonates (SMD ‐0.59, 95% CI ‐0.92 to ‐0.26; I² = 87%), based on very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias), once for imprecision, and once for inconsistency (see Table 4).

7.2. Analysis.

Comparison 7: Facilitated tucking X preterm neonate, Outcome 2: Immediate pain regulation

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

One study with 60 participants (treatment = 30, control = 30), included qualitatively, demonstrated that facilitated tucking may be efficacious at reducing pain reactivity during vaccinations in full‐term born neonates (Kucukoglu 2015), based on low‐certainty evidence downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

One study with two treatment arms and 47 participants (treatment = 32, control = 15), included qualitatively, concluded that facilitated tucking by a mother in a sitting‐up position may reduce pain reactivity (Khasanah 2019). This was based on very low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and twice for imprecision (once since only one study reported on this outcome and once because the treatment n of this study in the two treatment arms did not exceed the threshold for imprecision of 30x). 

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

Of the 10 studies reporting on this outcome, only one described an adverse event. Axelin 2006 reported that one participant developed septicaemia after receiving the experimental care, whereas the other nine studies did not observe any adverse events. We deemed this to be very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias) and once for imprecision.

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

No studies reported on this outcome.

Summary of treatment effects: facilitated tucking

Preterm born neonates

Very low‐certainty evidence suggests that facilitated tucking may be efficacious at reducing pain reactivity and improving immediate pain regulation in preterm born neonates. One study reported one adverse event (septicaemia) following this intervention. 

Full‐term born neonates

Based on low‐certainty evidence, one study included qualitatively suggests that facilitated tucking may be an efficacious treatment for full‐term neonates (Kucukoglu 2015). 

Full‐term born older infants

An included study described qualitatively concluded that facilitated tucking by the mother may reduce pain reactivity (Khasanah 2019). 

6. Touch/massage‐related interventions versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Three studies (four treatment arms) analysed the effect of touch‐related interventions on pain reactivity in preterm born neonates (see Analysis 8.1). The total number of analysed participants was 230 (control = 115, treatment = 115). The mean sample size across studies was 44 (SD = 19; range = 32 to 65). Three of the four treatment arms were cross‐over. One study reported two groups separately, which both received touch/massage with separate treatment groups (Chik 2012). We entered these two groups into the analysis as separate studies since they used different participants. Touch‐related interventions may be efficacious in reducing pain reactivity in preterm neonates (SMD ‐0.92, 95% CI ‐1.60 to ‐0.25; I² = 91%), based on very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias), once for inconsistency, and once for imprecision.

8.1. Analysis.

Comparison 8: Touch/massage related X preterm neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Two studies analysed whether touch‐related interventions were efficacious for immediate pain regulation in preterm born neonates (see Analysis 8.2). The total number of analysed participants was 112 (control = 56, treatment = 56). The mean sample size across studies was 45 (SD = 30; range = 23 to 66). One study was between‐group and one was cross‐over. Touch‐related interventions may be efficacious for improving the immediate pain regulation of preterm neonates (SMD ‐1.24, 95% CI ‐1.85 to ‐0.63; I² = 67%), based on evidence of low certainty downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

8.2. Analysis.

Comparison 8: Touch/massage related X preterm neonate, Outcome 2: Immediate pain regulation

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Six studies (seven treatment arms) analysed whether touch‐related interventions were efficacious on pain reactivity in full‐term born neonates (see Analysis 9.1). One study reported two groups separately, which both received different types of touch‐related interventions (acupressure versus foot massage) (Ozkan 2019). Thus, we entered these two intervention groups into the analysis as separate studies since they used different participants, and we entered the control group only once to avoid double‐counting. The total number of analysed participants was 520 (control = 236, treatment = 284). The mean sample size across studies was 73 (SD = 41; range = 17 to 139). Four of the six studies were between‐group. Touch‐related interventions may be efficacious for reducing pain reactivity in full‐term born neonates (SMD ‐0.81, 95% CI ‐1.22 to ‐0.40; I² = 83%), based on low‐certainty evidence downgraded twice for very serious study limitations due to overall risk of bias.

9.1. Analysis.

Comparison 9: Touch/massage related X full‐term neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Three studies (four treatment arms) analysed the effect of touch on the immediate pain regulation of full‐term born neonates (see Analysis 9.2). One study reported two groups separately, which both received different types of touch‐related interventions (acupressure versus foot massage) (Ozkan 2019). Thus, we entered these two intervention groups into the analysis as separate studies since they used different participants, and we entered the control group only once to avoid double‐counting. The total number of analysed participants was 295 (control = 124, treatment = 171). The mean sample size across studies was 82 (SD = 45; range = 32 to 139). All treatment arms were between‐group studies. Touch‐related interventions may be efficacious in improving immediate pain regulation in full‐term born neonates (SMD ‐0.92, 95% CI ‐1.57 to ‐0.28; I² = 85%), based on very low‐certainty evidence downgraded twice for very serious study limitations (overall risk of bias) and once for inconsistency.

9.2. Analysis.

Comparison 9: Touch/massage related X full‐term neonate, Outcome 2: Immediate pain regulation

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Five studies analysed whether touch‐related interventions were efficacious on pain reactivity in full‐term born older infants (see Analysis 10.1). The total number of participants was 368 (control = 184, treatment = 184). The mean sample size across studies was 74 (SD = 26; range = 60 to 120). All the studies were between‐group. Analyses showed that touch‐related interventions may be efficacious for reducing pain reactivity in older infants (SMD ‐1.28, 95% CI ‐1.91 to ‐0.65; I² = 87%), based on low‐certainty evidence, which we downgraded once for serious study limitations (risk of bias) and once for inconsistency. We described one included study qualitatively due to inconsistent reporting of statistics precluding its inclusion in the meta‐analysis (Karaca 2016). Karaca 2016 (70 participants; control = 35, treatment = 35) concluded that a touch‐related intervention (flicking the vaccination site) may reduce pain reactivity in older infants, based on low‐certainty evidence downgraded once for serious study limitations (overall risk of bias) and once for imprecision. 

10.1. Analysis.

Comparison 10: Touch/massage related X older infant, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Three studies analysed the effect of touch on the immediate pain regulation of full‐term born older infants (see Analysis 10.2). The total number of participants was 244 (control = 122, treatment = 122). The mean sample size across studies was 81 (SD = 34; range = 60 to 120). All treatment arms were between‐group studies. The results indicated that touch‐related interventions may be efficacious at improving immediate pain regulation in older infants (SMD ‐1.45, 95% CI ‐2.04 to ‐0.86; I² = 75%), based on very low‐certainty evidence, which we downgraded twice for very serious study limitations (risk of bias) and once for inconsistency. Karaca 2016 (70 participants; control = 35, treatment = 35), included qualitatively, concluded that a touch‐related intervention (flicking the vaccination site) may improve immediate pain regulation in older infants, based on low‐certainty evidence downgraded once for serious study limitations (overall risk of bias) and once for imprecision. 

10.2. Analysis.

Comparison 10: Touch/massage related X older infant, Outcome 2: Immediate pain regulation

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

Only one study reported on this outcome, and it did not find any adverse events. We deemed this to be evidence of moderate certainty, downgraded once for imprecision. 

Full‐term born neonates (infants born at 37 weeks until one month of age)

Five studies reporting on this outcome did not find any adverse effects. We deemed this to be low‐certainty evidence, which we downgraded twice for very serious study limitations due to overall risk of bias.

Full‐term born older infants (over one month to 36 months of age)

Six studies, which explicitly reported on adverse events, did not observe any during the studies. We deemed this to be low‐certainty evidence, which we downgraded twice for very serious study limitations due to overall risk of bias.

Summary of treatment effects: touch/massage‐related interventions

Preterm born neonates

The findings from both the meta‐analyses and one qualitatively described study suggest that touch‐related interventions may be efficacious in reducing pain reactivity and improving immediate pain regulation in preterm born neonates (low‐ to very low‐certainty evidence). The studies did not report any adverse events.

Full‐term born neonates

Low‐certainty evidence suggests that touch interventions may be efficacious for reducing pain reactivity in full‐term neonates. Very low‐certainty evidence suggests that touch interventions may also be effective at improving immediate pain regulation in full‐term neonates. The studies did not report any adverse events.

Full‐term born older infants

Very low‐ to low‐certainty evidence suggests that touch may be efficacious in reducing pain reactivity and immediate pain regulation scores in older infants. The studies did not report any adverse events.

7. Toy distraction versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Four studies with five treatment arms analysed the effects of toy distraction on the pain reactivity of older infants (see Analysis 11.1). In these studies, infants were shown a toy (i.e. rattle, spin wheel) and encouraged to direct their attention towards it using verbal cues (i.e. 'Look at this!'). The total number of analysed participants was 293 (control = 131, treatment = 162). The mean sample size across studies was 59 (SD = 24; range = 32 to 81). All treatment arms were between‐group. One study had two separate treatment arms, both of which involved toy distraction, and we included them as separate studies (Hillgrove Stuart 2013). Overall, there is evidence that toy distraction is not efficacious in reducing pain reactivity in older infants (SMD ‐0.15, 95% CI ‐0.39 to 0.08; I² = 0%), based on moderate‐certainty evidence, which we downgraded once for serious study limitations due to risk of bias. We described one included study qualitatively due to inconsistent reporting of statistics precluding its inclusion in the meta‐analysis (Dabas 2019). Dabas 2019 (100 participants; treatment = 75 across three treatment arms, control = 25) concluded that toy distraction may be effective in reducing pain reactivity, based on low‐certainty evidence downgraded once for serious study limitations (overall risk of bias) and once for imprecision. 

11.1. Analysis.

Comparison 11: Toy distraction x older infant, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

One study analysed the effect of toy distraction on the immediate pain regulation of older infants with two treatment arms, both of which involved toy distraction; we included them as separate studies (Hillgrove Stuart 2013) (see Analysis 11.2). The total number of analysed participants was 99 (control = 34, treatment = 65) with two treatment arms. Toy distraction was not efficacious in improving immediate pain regulation for older infants (SMD ‐0.08, 95% CI ‐0.50 to 0.34; I² = 0%), based on low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

11.2. Analysis.

Comparison 11: Toy distraction x older infant, Outcome 2: Immediate pain regulation

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Three studies reporting on this outcome did not find any adverse effects. We deemed this to be moderate‐certainty evidence, downgraded once for serious study limitations due to overall risk of bias.

Summary of treatment effects: toy distraction

Preterm born neonates

No studies reported on this outcome.

Full‐term born neonates

No studies reported on this outcome.

Full‐term born older infants

Low‐ to moderate‐certainty evidence suggests that toy distraction is not more efficacious than a no‐treatment control for reducing pain reactivity and immediate pain regulation in older infants. No studies reported any adverse events.

8. Structured parent involvement versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Six studies with seven treatment arms analysed the effect of structured parent involvement (see Analysis 12.1). One study reported two groups of infants (six and 18 months old infants) separately, which both received structured parent involvement with separate control groups (Pillai Riddell 2018). We entered these two groups into the analysis as separate studies since they used different participants. Structured parent involvement included studies whereby parents were given specific instructions on what to do with their infant during a painful procedure (either verbally/verbal coaching or via an information sheet). The total number of analysed participants was 653 (control = 329, treatment = 324). The mean sample size across studies was 109 (SD = 46; range = 50 to 160). All treatment arms were between‐group. The overall pooled effect indicates that there is evidence that structured parent involvement was no more efficacious than a no‐treatment control group in reducing pain reactivity in older infants (SMD ‐0.18, 95% CI ‐0.40 to 0.03; I² = 46%), based on moderate‐certainty evidence, which we downgraded once for serious study limitations due to overall risk of bias (see Table 5). One study included qualitatively (177 participants; control = 88, treatment = 89) analysed the long‐term benefits of consistent pain management via structured parent involvement on pain reactivity at 15 months of age (Taddio 2017a). This study did not find a long‐term benefit of structured parent involvement during infant vaccinations on future infant pain reactivity at 15 months, based on moderate‐certainty evidence downgraded once for imprecision as it was the only study examining the long‐term effects of this intervention. 

12.1. Analysis.

Comparison 12: Structured parent involvement X older infant, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Four studies with five treatment arms analysed the efficacy of structured parent involvement on immediate pain regulation for older infants (see Analysis 12.2). The total number of analysed participants was 728 (control = 356, treatment = 372). The mean sample size across studies was 182 (SD = 69; range = 128 to 284). All treatment arms were between‐groups. The intervention was not more efficacious than a no‐treatment control in improving pain regulation for older infants (SMD ‐0.09, 95% CI ‐0.40 to 0.21; I² = 74%), based on low‐certainty evidence downgraded once for serious study limitations (overall risk of bias) and once for inconsistency (see Table 5). It should be noted that a pattern emerged such that structured parent involvement showed efficacy in older age groups of infants (i.e. 18 months), as opposed to younger (i.e. six to eight months); however, further research on this intervention across ages is warranted. One study included qualitatively (177 participants; control = 88, treatment = 89) analysed the long‐term benefits of consistent pain management via structured parent involvement on immediate pain regulation at 15 months of age (Taddio 2017a). This study did not find a long‐term benefit of structured parent involvement during infant vaccinations on future infant pain regulation at 15 months, based on moderate‐certainty evidence downgraded once for imprecision as it was the only study examining the long‐term effects of this intervention. 

12.2. Analysis.

Comparison 12: Structured parent involvement X older infant, Outcome 2: Immediate pain regulation

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Of the six studies that reported on adverse effects, five found no adverse effects. Taddio 2017 offered parents an active educational video about infant soothing, and one of the pain management strategies included in the video led to some mild adverse effects. Specifically, the lidocaine cream led to transient skin reactions (e.g. pallor/redness) that were common across participants (pallor ranging from 32.1% to 71.8% and redness ranging from 58.5% to 80.2% across participants). We deemed this to be low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for inconsistency (see Table 5).

Summary of treatment effects: structured parent involvement

Preterm born neonates

No studies reported on this outcome.

Full‐term born neonates

No studies reported on this outcome.

Full‐term born older infants

Based on low‐ to moderate‐certainty evidence, structured parent involvement is not efficacious for pain reactivity or pain regulation. One study reported transient skin reactions resulting from one of the strategies (lidocaine cream) endorsed in a parent educational video. 

9. Sound reduction versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Two studies analysed the effect of sound reduction on pain reactivity in preterm born neonates (see Analysis 13.1). The total number of analysed participants was 144 (control = 72, treatment = 72). The mean sample size across studies was 72 (SD = 57; range = 32 to 112). Both studies were between‐group design. Sound reduction was not found to be efficacious in reducing pain reactivity in preterm born neonates (SMD ‐0.64, 95% CI ‐2.06 to 0.78; I² = 91%), based on very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias) and once for inconsistency. 

13.1. Analysis.

Comparison 13: Sound reduction X preterm neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Only one included study inspected the effects of sound reduction on immediate pain regulation in preterms; thus, we described it qualitatively. Ayazi 2017 (112 participants; treatment = 56, control = 56) concluded that ear protectors were efficacious at improving immediate pain regulation up to five minutes after a venipuncture. This was based on very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias) and once for imprecision. 

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

No studies reported on this outcome.

Summary of treatment effects: sound reduction

Preterm born neonates

Based on very low‐certainty data, sound reduction is not an efficacious intervention for reducing pain reactivity or improving immediate pain regulation in preterm born neonates.

Full‐term born neonates

No studies reported on this outcome.

Full‐term born older infants

No studies reported on this outcome.

10. Sound addition versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Seven studies (with eight treatment arms) analysed the effect of sound addition (e.g. exposing the infant to soothing sounds during the procedure, such as a recording of the mother's voice or intrauterine sounds) on pain reactivity in preterm born neonates (see Analysis 14.1). One study had two separate treatment arms, both of which involved exposure to distinct soothing sounds (white noise and maternal voice) and we included them as separate studies (Kahraman 2020). The total number of analysed participants was 362 (control = 174, treatment = 188). The mean sample size across studies was 49 (SD = 18; range = 19 to 72). The majority of treatment arms (7/8) were between‐group. The results found that exposing preterm born neonates to soothing sounds may reduce pain reactivity (SMD ‐0.89, 95% CI ‐1.43 to ‐0.34; I² = 86%), based on very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias), once for inconsistency, and once for imprecision.

14.1. Analysis.

Comparison 14: Sound addition X preterm neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Four studies analysed the effect of exposure to soothing sounds on the immediate pain regulation of preterm neonates (see Analysis 14.2), with a sample size of 360 (control = 182, treatment = 178). All studies were between‐group with a mean sample size of 90 (SD = 71; range = 40 to 196). This intervention was not efficacious in improving immediate pain regulation of preterm born neonates (SMD ‐0.92, 95% CI ‐2.20 to 0.36; I² = 96%), based on low‐certainty evidence downgraded once for serious study limitations (overall risk of bias) and once for inconsistency.

14.2. Analysis.

Comparison 14: Sound addition X preterm neonate, Outcome 2: Immediate pain regulation

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Two studies investigated the effect of exposure to soothing sounds on pain reactivity in full‐term born neonates; however, they could not be combined in a meta‐analysis as one of the studies did not report the necessary statistics for inclusion in the analysis (Azarmnejad 2015). Therefore, the two included studies are separately described in a qualitative manner. One study (60 participants; treatment = 30, control = 30) concluded that exposing full‐term neonates to white noise may reduce their pain reactivity (Cetinkaya 2020). Azarmnejad 2015 (30 participants; treatment = 15, control = 15) concluded that exposure to soothing maternal voice may also reduce pain reactivity in full‐term neonates. This was based on very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias) and once for imprecision.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

One study with 60 participants (treatment = 30, control = 30), included qualitatively, demonstrated that exposing older infants to their mothers' voices may reduce their pain reactivity (Erdogan 2020). This was based on low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

One study with 60 participants (treatment = 30, control = 30), included qualitatively, demonstrated that exposing older infants to their mothers' voices may improve their immediate pain regulation (Erdogan 2020). This was based on low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

Two studies reporting on this outcome did not find any adverse effects. We deemed this to be very low‐certainty evidence, downgraded twice for very serious study limitations (overall risk of bias) and once for imprecision.

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

No studies reported on this outcome.

Summary of treatment effects: sound addition

Preterm born neonates

The pooled results suggest that exposure to soothing sounds may be efficacious for reducing pain reactivity, but not improving immediate pain regulation, based on very low‐ to low‐certainty evidence. Studies did not report any adverse events.

Full‐term born neonates

Based on very low‐certainty evidence, two studies reported that sound addition was efficacious at reducing pain reactivity in full‐term neonates.

Full‐term born older infants

Based on low‐certainty evidence, one study reported that sound addition may be efficacious at reducing pain reactivity and improving immediate pain regulation in older infants.

11. Smell addition versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Five studies with eight treatment arms analysed the effect of exposure to soothing smells (e.g. mother's, vanilla, lavender, and breast milk smells) on pain reactivity in preterm born neonates (see Analysis 15.1). One study had two separate treatment arms, both of which involved exposure to different soothing smells (breast milk versus vanilla smell) and we included them as separate studies (Jebreili 2015). Another study had three separate treatment arms, all of which involved exposure to different soothing smells (amniotic fluid, breast milk, and mothers' smells) and we also included them as separate studies (Alemdar 2017a). All treatment arms were between‐group. The total number of analysed participants was 383 (control = 150, treatment = 233). The mean sample size across studies was 76 (SD = 36; range = 61 to 135). Exposing preterm neonates to soothing smells may be efficacious in reducing pain reactivity (SMD ‐0.41, 95% CI ‐0.72 to ‐0.10; I² = 49%), based on very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias) and once for imprecision.

15.1. Analysis.

Comparison 15: Smell addition X preterm neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Five studies with eight treatment arms analysed the effect of exposure to soothing smells (e.g. mother's smell, vanilla, lavender, breast milk) on immediate pain regulation in preterm neonates (see Analysis 15.2). One study had two separate treatment arms, both of which involved exposure to different soothing smells (breast milk versus vanilla smell) and we included them as separate studies (Jebreili 2015). Another study had three separate treatment arms, all of which involved exposure to different soothing smells (amniotic fluid, breast milk, and mothers' smells) and we also included them as separate studies (Alemdar 2017a). All treatment arms were between‐group. The total number of analysed participants was 383 (control = 150, treatment = 233). The mean sample size across studies was 76 (SD = 36; range = 61 to 135). The studies found that exposing preterm neonates to soothing smells may be efficacious in improving immediate pain regulation (SMD ‐0.57, 95% CI ‐0.88 to ‐0.27; I² = 47%), based on very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias) and once for imprecision.

15.2. Analysis.

Comparison 15: Smell addition X preterm neonate, Outcome 2: Immediate pain regulation

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Two studies (one with three treatment arms; Akcan 2016) analysed the effect of exposure to soothing smells on pain reactivity in full‐term born neonates (see Analysis 16.1). All treatment arms were between‐group. The three treatment arms involved exposure to distinct smells, and we included them as separate studies. The total number of analysed participants was 182 (control = 65, treatment = 117) with a mean sample size across studies of 91 (SD = 16; range = 80 to 102). Exposing full‐term neonates to soothing smells may reduce pain reactivity (SMD ‐0.77, 95% CI ‐1.09 to ‐0.44; I² = 0%), based on very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias) and once for imprecision.

16.1. Analysis.

Comparison 16: Smell addition X full‐term neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Two studies with five treatment arms analysed the effect of exposure to soothing smells on immediate pain regulation in full‐term neonates (see Analysis 16.2). All treatment arms were between‐group. One study had two separate treatment arms, both of which involved exposure to different soothing smells, and we included them as separate studies (Sadathosseini 2013). Another study had three separate treatment arms, all of which involved exposure to different soothing smells, and we also included them as separate studies (Akcan 2016). The total number of analysed participants was 237 (control = 70, treatment = 167). The mean sample size across studies was 119 (SD = 23; range = 102 to 135). The results indicated that exposing full‐term neonates to soothing smells may improve immediate pain regulation (SMD ‐0.41, 95% CI ‐0.80 to ‐0.01; I² = 46%), based on moderate‐certainty evidence, which we downgraded once for serious study limitations due to overall risk of bias. Interestingly, Sadathosseini 2013 exposed full‐term neonates to both familiar and unfamiliar smells. Only exposure to familiar smells demonstrated evidence for an analgesic effect of smell addition compared to no treatment.

16.2. Analysis.

Comparison 16: Smell addition X full‐term neonate, Outcome 2: Immediate pain regulation

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

Two studies reported on this outcome and did not find any adverse effects. We deemed this to be moderate‐certainty evidence, which we downgraded once for serious study limitations due to the overall risk of bias.

Full‐term born neonates (infants born at 37 weeks until one month of age)

One study reported on this outcome, and it did not find any adverse effects. We deemed this to be low‐certainty evidence downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Full‐term born older infants (over one month to 36 months of age)

No studies reported on this outcome.

Summary of treatment effects: smell addition

Preterm born neonates

The findings from this review suggest that exposure to soothing smells may be efficacious in improving pain reactivity and immediate pain regulation (very low‐certainty evidence). No adverse events were reported.

Full‐term born neonates

Very low‐certainty evidence suggests that exposure to soothing smells may reduce pain reactivity in full‐term neonates. Moderate‐certainty evidence suggests that exposure to soothing smells may also improve immediate pain regulation in full‐term neonates. No adverse events were reported.

Full‐term born older infants

No studies reported on this outcome.

12. Light reduction versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Two studies analysed the effect of light reduction (e.g. limiting the amount of light infants are exposed to by covering their eyes or their incubator) on pain reactivity in preterm born neonates (see Analysis 17.1). The total number of analysed participants was 125 (control = 62, treatment = 63). The mean sample size across studies was 63 (SD = 1; range = 62 to 63). Both studies were between‐group. Reducing the amount of light that preterm neonates are exposed to may reduce pain reactivity (SMD ‐0.71, 95% CI ‐1.08 to ‐0.34; I² = 0%), based on evidence of moderate certainty, which we downgraded once for serious study limitations due to overall risk of bias. 

17.1. Analysis.

Comparison 17: Light reduction x preterm neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Two studies analysed the effect of light reduction on immediate pain regulation in preterm neonates (see Analysis 17.2). The total number of analysed participants was 125 (control = 62, treatment = 63). The mean sample size across studies was 63 (SD = 1; range = 62 to 63). Both studies were between‐group. The results found that reducing the amount of light that preterm neonates are exposed to may improve immediate pain regulation (SMD ‐1.16, 95% CI ‐1.53 to ‐0.78; I² = 0%), based on evidence of moderate certainty, which we downgraded once for serious study limitations due to overall risk of bias.

17.2. Analysis.

Comparison 17: Light reduction x preterm neonate, Outcome 2: Immediate pain regulation

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

No studies reported on this outcome.

Summary of treatment effects: light reduction

Preterm born neonates

The pooled results suggest that light reduction may be efficacious for reducing pain reactivity and improving immediate pain regulation, based on evidence of moderate certainty.

Full‐term born neonates

No studies reported on this outcome.

Full‐term born older infants

No studies reported on this outcome.

13. Heat addition versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

One study with 50 participants (treatment = 25, control = 25), included qualitatively, reported that heel warming may reduce pain reactivity in preterm neonates (Shu 2014), based on very low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and twice for imprecision. 

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Three studies analysed the effect of heat addition (e.g. heel warming) on pain reactivity in full‐term born neonates (see Analysis 18.1). The total number of analysed participants was 232 (control = 116, treatment = 116). All studies were between‐group. The mean sample size across studies was 77 (SD = 24; range = 52 to 100). The pooled results suggest that heat addition is not efficacious in reducing pain reactivity (SMD ‐0.12, 95% CI ‐0.42 to 0.17; I² = 22%), based on low‐certainty evidence, which we downgraded twice for very serious study limitations due to high risk of bias.

18.1. Analysis.

Comparison 18: Heat addition X full‐term neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

Shu 2014 did not report any adverse effects during heel warming, based on very low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and twice for imprecision.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Only one study explicitly reported this outcome, and it did not find any adverse effects as a result of heel warming (Mir 2018). We deemed this to be very low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and twice for imprecision.

Full‐term born older infants (over one month to 36 months of age)

No studies reported on this outcome.

Summary of treatment effects: heat addition

Preterm born neonates

Based on very low‐certainty evidence, one study qualitatively concluded that heat addition was efficacious at reducing pain reactivity in preterm neonates. This study did not report any adverse events.

Full‐term born neonates

The pooled results suggest that heat addition was not efficacious for reducing pain reactivity, based on evidence of low certainty. Studies did not find any adverse events following this intervention. 

Full‐term born older infants

No studies reported on this outcome.

14. Cold addition versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Two studies analysed the effect of cold addition on pain reactivity in full‐term born neonates (see Analysis 19.1). The total number of analysed participants was 142 (control = 71, treatment = 71). Both studies were between‐group. The mean sample size across studies was 71 (SD = 10; range = 64 to 78). The pooled results suggest that cooling the procedure site may reduce pain reactivity (SMD ‐0.85, 95% CI ‐1.48 to ‐0.23; I² = 69%), based on low‐certainty evidence, which we downgraded once for very serious study limitations due to overall risk of bias and once for inconsistency.

19.1. Analysis.

Comparison 19: Cold addition X full‐term neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

One study with 78 participants (treatment = 39, control = 39), included qualitatively, reported that cooling may improve immediate pain regulation (Najafi 2019), based on low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for imprecision. 

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

No studies reported on this outcome.

Summary of treatment effects: cold addition

Preterm born neonates

No studies reported on this outcome.

Full‐term born neonates

The pooled results suggest that cooling may reduce pain reactivity, based on evidence of low certainty. One study reported that cooling may also improve immediate pain regulation (low‐certainty evidence). 

Full‐term born older infants

No studies reported on this outcome.

15. Multisensory bundle versus no‐treatment control

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Five studies analysed the effect of multisensory bundles (i.e. combinations of interventions of different modalities) on pain reactivity in preterm born neonates (see Analysis 20.1). The total number of analysed participants was 280 (control = 138, treatment = 142). The mean sample size across studies was 40 (SD = 26; range = 14 to 80). Three of the five studies were cross‐over. Multisensory bundles were found to potentially be efficacious in reducing pain reactivity (SMD ‐2.70, 95% CI ‐4.16 to ‐1.23; I² = 97%), based on evidence of very low certainty downgraded once for serious study limitations (overall risk of bias), once for inconsistency, and once for imprecision.

20.1. Analysis.

Comparison 20: Multisensory bundle X preterm neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Four studies analysed the effect of multisensory bundles on immediate pain regulation in preterm born neonates (see Analysis 20.2). The total number of analysed participants was 228 (control = 114, treatment = 114). The mean sample size across studies was 42 (SD = 28; range = 14 to 80). Half of the studies were cross‐over. Multisensory bundles may be efficacious in improving immediate pain regulation (SMD ‐1.49, 95% CI ‐2.92 to ‐0.06; I²= 89%), based on evidence of very low certainty, downgraded once for serious study limitations (overall risk of bias), once for inconsistency, and once for imprecision.

20.2. Analysis.

Comparison 20: Multisensory bundle X preterm neonate, Outcome 2: Immediate pain regulation

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Two studies analysed the effect of multisensory bundles (i.e. combinations of interventions of different modalities) on pain reactivity in full‐term neonates (see Analysis 21.1). The total number of analysed participants was 227 (control = 119, treatment = 108). The mean sample size across studies was 114 (SD = 104; range = 40 to 187). Both studies were between‐group. Multisensory bundles had no effect on reducing pain reactivity (SMD 0.00, 95% CI ‐0.29 to 0.30; I² = 9%), based on evidence of moderate certainty, downgraded once for serious study limitations (overall risk of bias).

21.1. Analysis.

Comparison 21: Multisensory bundle X full‐term neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

One study with 60 participants (treatment = 30, control = 30), included qualitatively, reported that a multisensory bundle (massage and cooling) may reduce pain reactivity in older infants (Khosravan 2018), based on low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for imprecision. 

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

One study with 114 participants (treatment = 58, control = 56), included qualitatively, reported that a multisensory bundle based on the five S's (swaddling, side/stomach position, shushing, swinging, and sucking) may improve immediate pain regulation (Harrington 2012), based on low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for imprecision. 

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

Four studies reported on this outcome, and they did not uncover any adverse effects relating to multisensory bundles. We deemed this to be low‐certainty evidence, downgraded twice for very serious study limitations due to the high risk of bias.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Only one study explicitly reported this outcome, and it did not find any adverse effects (Bellieni 2002). We deemed this to be very low‐certainty evidence which we downgraded once for serious study limitations (overall risk of bias) and twice for imprecision.

Full‐term born older infants (over one month to 36 months of age)

One study explicitly reported this outcome, and it did not find any adverse effects (Khosravan 2018). We deemed this to be low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Summary of treatment effects: multisensory bundle

Preterm born neonates

The pooled results suggest that multisensory bundles may be efficacious for reducing pain reactivity and improving immediate pain regulation, based on evidence of very low certainty. No trials reported any adverse events. 

Full‐term born neonates

The pooled results suggest that multisensory bundles may have no effect on reducing pain reactivity and improving immediate pain regulation in full‐term neonates, based on evidence of moderate certainty. No trials reported any adverse events. 

Full‐term born older infants

Studies described qualitatively demonstrated that multisensory bundles may reduce pain behaviours in older infants, based on uncertain evidence. Studies did not report adverse events. 

16. Additive non‐pharmacological intervention on sweet solution versus sweet solution only

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Seven studies analysed the additive effect of non‐pharmacological interventions on a sweet solution, such as glucose or sucrose, on pain reactivity in preterm born neonates (see Analysis 22.1). The total number of analysed participants was 456 (control = 221, treatment = 235). The mean sample size across studies was 62 (SD = 38; range = 24 to 124). Six of the seven studies were between‐group. Six studies administered sucrose, ranging from 20% to 30%, while one study administered 30% dextrose. Non‐pharmacological interventions in addition to sweet solutions may be more efficacious than sweet solutions alone in reducing pain reactivity (SMD ‐1.10, 95% CI ‐1.96 to ‐0.25; I² = 94%), based on evidence of low certainty, which we downgraded once for serious study limitations (overall risk of bias) and once for inconsistency. Two studies demonstrated notably stronger effects than the rest, with both studies examining the additive effect of non‐nutritive sucking (NNS) on sucrose (Gao 2018; Kristoffersen 2011). Therefore, it may be that administering sucrose with non‐nutritive sucking may be a particularly efficacious intervention as opposed to sucrose alone in preterm born neonates; however, more research is warranted in this area. 

22.1. Analysis.

Comparison 22: Additive non‐pharmacological intervention on sweet solution X preterm neonate, Outcome 1: Pain reactivity

We also described two included studies qualitatively in this section as these studies did not report the required statistics for inclusion in the meta‐analysis. Petrillo 2010 (10 participants; control = 5, treatment = 5) reported that a multisensory bundle (non‐nutritive sucking, containment, sound addition) in addition to sucrose may be more effective than sucrose alone at reducing pain reactivity in preterms. Ti 2015 (44 participants; control = 23, treatment = 21) reported that facilitated tucking and non‐nutritive sucking in addition to sucrose may be more effective than sucrose alone at reducing pain reactivity. This qualitatively described finding is based on uncertain evidence, downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Four studies analysed the additive effect of non‐pharmacological interventions on a sweet solution on immediate pain regulation in preterm born neonates (see Analysis 22.2). The total number of analysed participants was 411 (control = 197, treatment = 214). The mean sample size across studies was 103 (SD = 75; range = 43 to 200). All studies were between‐group. Three studies administered sucrose, ranging from 20% to 24%, and one study administered a 25% glucose solution. Non‐pharmacological interventions in addition to sweet solutions were not more efficacious than sweet solutions alone in improving immediate pain regulation (SMD ‐0.90, 95% CI ‐2.00 to 0.20; I² = 96%), based on evidence of low certainty, which we downgraded once for serious study limitations (overall risk of bias) and once for inconsistency.

22.2. Analysis.

Comparison 22: Additive non‐pharmacological intervention on sweet solution X preterm neonate, Outcome 2: Immediate pain regulation

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Four studies analysed the additive effect of non‐pharmacological interventions on a sweet solution, such as glucose or sucrose, on pain reactivity in full‐term born neonates (see Analysis 23.1). The total number of analysed participants was 323 (control = 128, treatment = 135). The mean sample size across studies was 66 (SD = 22; range = 40 to 90). All the studies were between‐group. Three of the four studies administered sucrose and one administered glucose solutions (ranging from 20% to 30%). Non‐pharmacological interventions in addition to a sweet solution may be more efficacious than the sweet solution alone in reducing pain reactivity (SMD ‐1.26, 95% CI ‐2.45 to ‐0.07; I² = 94%), based on evidence of low certainty, which we downgraded once for serious study limitations (overall risk of bias) and once for inconsistency. One study demonstrated a notably stronger effect than the rest, and it analysed the additive effect of non‐nutritive sucking on sucrose (with all infants also placed in warmers) (Thakkar 2016). 

23.1. Analysis.

Comparison 23: Additive non‐pharmacological intervention on sweet solution X full‐term neonate, Outcome 1: Pain reactivity

We also describe one study qualitatively as it did not meet the eligibility criteria for inclusion in the meta‐analysis. Petrillo 2010 (75 participants; control = 25, treatment = 50 across two treatment arms) reported that a multisensory bundle (non‐nutritive sucking, containment, sound addition) in addition to sucrose may be more effective than sucrose alone at reducing pain reactivity in full‐term neonates, based on uncertain evidence, downgraded once for serious study limitations (overall risk of bias) and once for imprecision. 

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Five studies with six treatment arms analysed the additive effect of non‐pharmacological interventions on a sweet solution on immediate pain regulation in full‐term born neonates (see Analysis 23.2). One study reported two groups separately, which both received non‐pharmacological interventions in addition to a sweet solution with separate control groups (Leng 2016). We entered these two groups into the analysis as separate studies since they used different participants. The total number of analysed participants was 875 (control = 442, treatment = 433). The mean sample size across studies was 175 (SD = 278; range = 29 to 671). All treatment arms were between‐group. All studies administered sucrose, ranging from 12% to 24%. Non‐pharmacological interventions in addition to sweet solutions may be more efficacious than sweet solutions alone in improving immediate pain regulation (SMD ‐0.53, 95% CI ‐0.93 to ‐0.13; I² = 84%), based on evidence of low certainty, which we downgraded once for serious study limitations (overall risk of bias) and once for inconsistency. Gray 2015, examining the additive effect of being placed in a warmer on sucrose, demonstrated a notably larger effect than the rest of the studies.

23.2. Analysis.

Comparison 23: Additive non‐pharmacological intervention on sweet solution X full‐term neonate, Outcome 2: Immediate pain regulation

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Two studies analysed the additive effect of non‐pharmacological interventions on a sweet solution on pain reactivity in term‐born older infants (see Analysis 24.1). The total number of analysed participants was 241 (control = 119, treatment = 122). The mean sample size across studies was 121 (SD = 1; range = 120 to 121). Both studies were between‐group and both administered 24% sucrose solutions. Non‐pharmacological interventions in addition to sucrose were not more efficacious than sucrose alone in reducing pain reactivity (SMD ‐0.14, 95% CI ‐0.39 to 0.10; I² = 0%), based on evidence of moderate certainty, which we downgraded once for serious study limitations due to overall risk of bias.

24.1. Analysis.

Comparison 24: Additive non‐pharmacological intervention on sweet solution X older infant, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

One study with 116 participants (treatment = 58, control = 58), included qualitatively, did not find evidence of a difference between administering a sweet solution only versus a sweet solution paired with a multisensory bundle of non‐pharmacological interventions (shushing, swaddling, sucking, side position, and swinging) for improving immediate pain regulation (Harrington 2012), based on low‐certainty evidence, downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

Four studies explicitly reported on this outcome. One of them reported that two participants in the intervention group exhibited adverse effects, such as vomiting and abdominal distension (Gao 2018). The other studies did not report any adverse events. We deemed this to be low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for inconsistency.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Of the seven studies that reported on this outcome, one found adverse effects. Thakkar 2016 described that one neonate in each of the intervention groups desaturated during the study. We deemed this to be very low‐certainty evidence, which we downgraded twice for very serious study limitations (overall risk of bias) and once for inconsistency.

Full‐term born older infants (over one month to 36 months of age)

Three studies reported on this outcome. Of the three, only Taddio 2014 reported adverse events, such as infant distress caused by sucrose administration. Although most infants settled, the administration of sucrose had to be aborted in four infants due to high distress. We deemed this to be low‐certainty evidence, downgraded once for serious study limitations (overall risk of bias) and once for inconsistency.

Summary of treatment effects: additive non‐pharmacological on a sweet solution

Preterm born neonates

Evidence of low certainty suggests that non‐pharmacological interventions in addition to sweet solutions may be more efficacious than sweet solutions alone in reducing pain reactivity, with the additive effect of non‐nutritive sucking on sucrose potentially being particularly efficacious in preterm born neonates. Non‐pharmacological interventions in addition to sweet solutions were not more efficacious than sweet solutions alone in improving immediate pain regulation, based on evidence of low certainty. One study reported the occurrence of two adverse events (vomiting and abdominal distension) related to this intervention. 

Full‐term born neonates

Evidence of low certainty suggests that non‐pharmacological interventions in addition to sweet solutions may be more efficacious than sweet solutions alone in reducing pain reactivity and improving immediate pain regulation. One study reported desaturation as an adverse event related to this intervention. 

Full‐term born older infants

Evidence of moderate certainty suggests that non‐pharmacological interventions in addition to sucrose were not more efficacious than sucrose alone in reducing pain reactivity. Based on low‐certainty evidence, a study described qualitatively did not find evidence for the additive effect of non‐pharmacological interventions on sweet solutions for improving immediate pain regulation. One study reported distress due to sucrose administration as an adverse effect.

17. Additive non‐pharmacological intervention on non‐nutritive sucking (NNS) versus NNS only

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Five studies with six treatment arms analysed the additive effect of non‐pharmacological interventions on non‐nutritive sucking on pain reactivity in preterm born neonates (see Analysis 25.1). The total number of analysed participants was 301 (control = 136, treatment = 165). The mean sample size across studies was 60 (SD = 23; range = 33 to 90). All studies were between‐group. One study had two separate treatment arms, both of which involved the addition of distinct non‐pharmacological interventions (gentle touch versus Yakson touch) on non‐nutritive sucking, and we analysed them as separate studies (Dur 2020). Other non‐pharmacological interventions in addition to non‐nutritive sucking may be more efficacious than non‐nutritive sucking alone in reducing pain reactivity (SMD ‐1.27, 95% CI ‐2.21 to ‐0.33; I² = 92%), based on evidence of very low certainty, which we downgraded once for serious study limitations (overall risk of bias), once for inconsistency, and once for imprecision. Both treatment arms in Dur 2020, which analysed the effects of touch‐related interventions on top of non‐nutritive sucking, demonstrated a notably stronger effect than the rest of the studies.

25.1. Analysis.

Comparison 25: Additive non‐pharmacological intervention on non‐nutritive sucking X preterm neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Three studies (four treatment arms) analysed the additive effect of non‐pharmacological interventions on non‐nutritive sucking on immediate pain regulation in preterm born neonates (see Analysis 25.2). The total number of analysed participants was 209 (control = 90, treatment = 119). One study had two separate treatment arms, both of which involved the addition of distinct non‐pharmacological interventions (gentle touch versus Yakson touch) on non‐nutritive sucking, and we analysed them as separate studies (Dur 2020). All studies were between‐group. The mean sample size across studies was 70 (SD = 23; range = 45 to 90). Other non‐pharmacological interventions in addition to non‐nutritive sucking may be more efficacious than non‐nutritive sucking alone in improving immediate pain regulation in preterm born neonates (SMD ‐1.32, 95% CI ‐2.59 to ‐0.05; I² = 94%), based on evidence of very low certainty, which we downgraded once for serious study limitations (overall risk of bias), once for inconsistency, and once for imprecision.

25.2. Analysis.

Comparison 25: Additive non‐pharmacological intervention on non‐nutritive sucking X preterm neonate, Outcome 2: Immediate pain regulation

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

One study with 77 participants (treatment = 40, control = 37), included qualitatively, studied the additive effects of noninvasive electrical stimulation of acupuncture points (NESAP) on top of NNS and facilitated tucking, demonstrating that NESAP does not have additive effect compared to NNS and facilitated tucking only for reducing pain reactivity (Mitchell 2016), based on low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for imprecision. 

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

One study with 77 participants (treatment = 40, control = 37), included qualitatively, studied the additive effects of NESAP on top of non‐nutritive sucking and facilitated tucking, demonstrating that NESAP does not have an additive effect compared to non‐nutritive sucking and facilitated tucking only for improving immediate pain regulation (Mitchell 2016), based on low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

Of the four studies that reported on this outcome, none found adverse effects. We deemed this to be low‐certainty evidence, downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Mitchell 2016 reported that no adverse effects were observed during the study. We deemed this to be low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Full‐term born older infants (over one month to 36 months of age)

No studies reported on this outcome.

Summary of treatment effects: additive non‐pharmacological on non‐nutritive sucking

Preterm born neonates

Evidence of very low certainty suggests that non‐pharmacological interventions in addition to non‐nutritive sucking may be more efficacious than non‐nutritive sucking alone in reducing pain reactivity or improving immediate pain regulation. No adverse events were reported following this intervention.

Full‐term born neonates

Based on low‐certainty evidence, one study included qualitatively reported that non‐pharmacological interventions in addition to non‐nutritive sucking were not more efficacious than non‐nutritive sucking alone in reducing pain reactivity or improving immediate pain regulation (Mitchell 2016). This study did not report any adverse events.

Full‐term born older infants

No studies reported on this outcome.

18. Additive non‐pharmacological intervention on swaddling versus swaddling only

Primary outcomes: pain response

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Two studies analysed the additive effect of non‐pharmacological interventions on swaddling on pain reactivity in full‐term born neonates (see Analysis 26.1). The total number of analysed participants was 180 (control = 90, treatment = 90). The mean sample size across studies was 90 (SD = 14; range = 80 to 100). Both studies were between‐group. Non‐pharmacological interventions in addition to swaddling were not more efficacious than swaddling alone in reducing pain reactivity (SMD ‐0.98, 95% CI ‐2.40 to 0.44; I² = 95%), based on evidence of very low certainty, which we downgraded twice for very serious study limitations (overall risk of bias) and once for inconsistency.

26.1. Analysis.

Comparison 26: Additive non‐pharmacological intervention on swaddling X full‐term neonate, Outcome 1: Pain reactivity

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Caglar 2017 (77 participants; treatment = 40, control = 37) also examined the additive effect of a ShotBlocker (device with contact points that is placed at the injection site and believed to provide counter‐stimulation to modulate the pain response) on top of swaddling (treatment arm included qualitatively), concluding that it did have an additive effect on top of swaddling at improving pain regulation in full‐term born neonates. This was based on low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

No studies reported on this outcome.

Summary of treatment effects: additive non‐pharmacological on swaddling

Preterm born neonates

No studies reported on this outcome.

Full‐term born neonates

Based on very low‐certainty evidence, the pooled results showed that adding non‐pharmacological interventions on top of swaddling was not more efficacious than swaddling only at reducing pain reactivity. One treatment arm included qualitatively concluded that adding non‐pharmacological interventions on top of swaddling may be more efficacious than swaddling only at improving immediate pain regulation (low‐certainty evidence) (Caglar 2017).

Full‐term born older infants

No studies reported on this outcome.

Interventions included and described qualitatively

1. Therapeutic touch

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

One study with 55 participants (treatment = 27, control = 28), included qualitatively, examined the effects of therapeutic touch (based on the theory of energy medicine, hands are held over the infant's body without actually touching the body) on pain reactivity, demonstrating that therapeutic touch does not reduce pain reactivity (Johnston 2013), based on very low‐certainty evidence, downgraded once for serious study limitations (overall risk of bias) and twice for imprecision. 

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

One study with 55 participants (treatment = 27, control = 28), included qualitatively, examined the effects of therapeutic touch (based on the theory of energy medicine, hands are held over the infant's body without actually touching the body) on pain regulation (Johnston 2013). It concluded that therapeutic touch does not improve immediate pain regulation based on very low‐certainty evidence, downgraded once for serious study limitations (overall risk of bias) and twice for imprecision. 

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

One study with 40 participants (treatment = 20, control = 20), included qualitatively, examined the effects of therapeutic touch (based on the theory of energy medicine, hands are held over the infant's body without actually touching the body) on pain reactivity (Kozub 2001), concluding that therapeutic touch was ineffective at reducing pain reactivity in older infants, based on very low‐certainty evidence, downgraded once for serious study limitations (overall risk of bias) and twice for imprecision. 

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Kozub 2001 reported no adverse effects during the study. We deemed this to be very low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and twice for imprecision.

Summary of treatment effects: therapeutic touch

Preterm born neonates

A study included qualitatively concluded that therapeutic touch is not an efficacious analgesic in preterm born neonates (Johnston 2013). 

Full‐term born neonates

No studies reported on this outcome.

Full‐term born older infants

One study included qualitatively concluded that therapeutic touch was ineffective at reducing pain behaviours in older infants (Kozub 2001). This study did not report any adverse events.

2. Vibration therapy

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

Two studies examined vibration therapy versus no‐treatment control in older infants but they could not be meta‐analysed as one did not provide the appropriate results to be analysed; thus these two studies were included qualitatively. One study with 60 participants (treatment = 30, control = 30) concluded that placing a vibration device at the venipuncture site was not efficacious at reducing pain reactivity (Secil 2014). The second study with 38 participants (treatment = 20, control = 18) demonstrated that vibration therapy had no effect on reducing pain reactivity in older infants (Benjamin 2016). These results are based on low‐certainty evidence, downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

Two studies examined vibration therapy versus no‐treatment control in older infants but they could not be meta‐analysed as one did not provide the appropriate results to be analysed; thus these two studies were included qualitatively. One study with 60 participants (treatment = 30, control = 30) concluded that placing a vibration device at the venipuncture site was not efficacious at improving immediate pain regulation (Secil 2014). The second study with 38 participants (treatment = 20, control = 18) demonstrated that vibration therapy had no effect on improving immediate pain regulation (Benjamin 2016). These results are based on low‐certainty evidence, downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

No studies reported on this outcome.

Summary of treatment effects: vibration therapy

Preterm born neonates

No studies reported on this outcome.

Full‐term born neonates

No studies reported on this outcome.

Full‐term born older infants

Two studies included qualitatively concluded that vibration therapy was ineffective at reducing pain behaviours in older infants based on uncertain evidence. 

3. Video distraction

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

One study with 90 participants (treatment = 49, control = 41), included qualitatively, examined the effects of video distraction (i.e. playing a DVD during the procedure) and demonstrated that it may reduce pain reactivity in older infants (Cohen 2002), based on uncertain evidence, downgraded once for serious study limitations (overall risk of bias) and once for imprecision. 

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

One study with 126 participants (treatment = 63, control = 63), included qualitatively, examined the effects of video distraction (i.e. playing a DVD during the procedure) and concluded that it may improve immediate pain regulation in older infants (Cohen 2006), based on uncertain evidence, downgraded once for serious study limitations (overall risk of bias) and once for imprecision. 

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

No studies reported on this outcome.

Summary of treatment effects: video distraction

Preterm born neonates

No studies reported on this outcome.

Full‐term born neonates

No studies reported on this outcome.

Full‐term born older infants

Based on two studies included qualitatively, it was concluded that video distraction may be efficacious at reducing pain behaviours in older infants.

4. Parent presence

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

One study with 20 participants (treatment = 10, control = 10), included qualitatively, examined the effects of parent presence during painful procedures and demonstrated that it has no effect on pain reactivity (Shaw 1982), based on very uncertain evidence, downgraded once for serious study limitations (overall risk of bias) and twice for imprecision. 

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

One study with 278 participants (treatment = 147, control = 131), included qualitatively, examined the effects of parent presence during painful procedures and demonstrated that it does not improve immediate pain regulation in older infants (Bauchner 1996), based on uncertain evidence, downgraded once for serious study limitations (overall risk of bias) and once for imprecision. 

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Of the two studies included qualitatively, only one reported on adverse events (Bauchner 1996). This study did not observe any adverse events following parent presence. We deemed this to be low‐certainty evidence, which we downgraded once for serious study limitations (overall risk of bias) and once for imprecision.

Summary of treatment effects: parent presence

Preterm born neonates

No studies reported on this outcome.

Full‐term born neonates

No studies reported on this outcome.

Full‐term born older infants

Based on two studies included qualitatively, parent presence was found to be ineffective at both reducing pain reactivity and improving pain regulation in older infants. No adverse events were reported. 

5. Co‐bedding

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

One study with 100 participants (treatment = 50, control = 50), included qualitatively, examined the effects of co‐bedding and demonstrated that co‐bedding does not reduce pain reactivity in preterm neonates (Badiee 2014), based on low‐certainty evidence, downgraded once for serious study limitations (overall risk of bias) and once for imprecision. 

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

Badiee 2014 reported no adverse events related to co‐bedding. We deemed this to be low‐certainty evidence, downgraded once for serious study limitations (overall risk of bias) and once for imprecision. 

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

No studies reported on this outcome.

Summary of treatment effects: co‐bedding

Preterm born neonates

One study included qualitatively concluded that co‐bedding does not reduce pain reactivity in preterms, based on uncertain evidence. This study did not report any adverse events.

Full‐term born neonates

No studies reported on this outcome.

Full‐term born older infants

No studies reported on this outcome.

6. Additive non‐pharmacological intervention on topical anaesthetic versus topical anaesthetic only

Preterm born neonates (born at 36 weeks gestation or less)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

Pain reactivity (zero to 30 seconds after the painful stimulus)

No studies reported on this outcome.

Immediate pain regulation (pain measurement closest to 30 seconds after the last painful stimulus)

One study with 24 participants (treatment = 10, control = 14), included qualitatively, examined the additive effects of a mechanical vibration device on top of a topical anaesthetic (LMX4 cream) (Bahorski 2015). It demonstrated that non‐pharmacological interventions combined with LMX4 cream may be more effective at improving immediate pain regulation compared to LMX4 cream only, based on very low‐certainty evidence, downgraded once for serious study limitations (overall risk of bias) and twice for imprecision. 

Secondary outcome: adverse events

Preterm born neonates (born at 36 weeks gestation or less)

No studies reported on this outcome.

Full‐term born neonates (infants born at 37 weeks until one month of age)

No studies reported on this outcome.

Full‐term born older infants (over one month to 36 months of age)

No studies reported on this outcome.

Summary of treatment effects: additive non‐pharmacological on topical anaesthetic

Preterm born neonates

No studies reported on this outcome.

Full‐term born neonates

No studies reported on this outcome.

Full‐term born older infants

One study included qualitatively concluded that a non‐pharmacological intervention (mechanical vibration) had an additive effect on top of a topical anaesthetic, and it effectively improved immediate pain regulation in older infants. 

Publication bias

In order to help minimise the influence of publication bias (i.e. the bias resulting from analysing only published studies), we used systematic methods to obtain studies that were not published. This included contacting researchers and clinicians broadly through emails and list‐serves, searching trial registration websites for relevant studies, and contacting specific researchers who presented data at relevant paediatric pain or pain conferences we attended. It is highly unlikely we were able to locate every relevant study, but we made a methodical attempt. We also plan to disseminate our findings broadly, by both co‐publishing parts of this review, and presenting it at conferences, with an invitation to researchers and clinicians who have worked on studies currently omitted to contact us, helping to further minimise this bias for future updates.

Discussion

The overall findings of the quantitative analysis are summarised succinctly in additional Table 6 with the explanation of the results presented in each of the summary of findings tables. Table 6 is only based on the 115 studies used for quantitative analyses. This table is the best summary of the quantitative analyses, including: analysis strategy (treatment x age group x pain phase), direction of effect (evidence supports efficacy; evidence supports inefficacy), and the confidence in the evidence using GRADE criteria. A notable caveat is that all analyses reflected that more research is needed to bolster our confidence in the direction of the findings, i.e. not one finding was based on a high‐certainty grade of the evidence. 

Summary of main results

Pain‐related reactivity and immediate pain‐related regulation

The results for all 138 included studies are offered below, separated by age group. 

For preterm born neonates, we analysed 12 different non‐pharmacological interventions and described an additional three in a qualitative manner. Of the analysed interventions, non‐nutritive sucking, swaddling, facilitated tucking, touch‐related, sound addition, smell addition, light reduction, multisensory bundling of interventions, and the additive effects of a non‐pharmacological intervention onto sweet solutions or onto non‐nutritive sucking may reduce pain‐related reactivity and/or improve immediate pain‐related regulation. However, the findings are mostly based on very low‐ to low‐certainty evidence. Of the three included interventions described qualitatively, heel warming may improve pain‐related reactivity, but this is again based on very low‐certainty evidence. Overall, the interventions based on the most substantial, and least heterogeneous, body of literature appear to be non‐nutritive sucking, swaddling, and facilitated tucking. Although the certainty of their evidence is very low, this body of literature suggests that these three interventions are the most promising for reducing pain behaviours in preterms (moderate to large effect sizes). 

For full‐term born neonates, we analysed 10 different non‐pharmacological interventions and described three in a qualitative manner. Although mostly based on a very low‐ to low‐certainty evidence base, analyses demonstrated that the following interventions may reduce pain‐related reactivity and/or improve immediate pain‐related regulation: non‐nutritive sucking, rocking/holding, swaddling, touch‐related, smell addition, cold addition, and the additive effects of a non‐pharmacological intervention onto sweet solutions. Of the three interventions described qualitatively, facilitated tucking and soothing sounds may also improve pain reactivity. However, with only one exception (smell addition ‐ moderate‐certainty), the findings are based on very low‐ to low‐certainty evidence with considerable heterogeneity. Overall, the intervention based on the most substantial and least heterogeneous body of literature appears to be non‐nutritive sucking. Although this intervention seems to show promise for reducing pain behaviours in full‐term newborns (large effect sizes), the evidence suggests very low certainty in the findings.

Finally, in full‐term born older infants, we analysed four different non‐pharmacological interventions and described 11 in a qualitative manner. Of the four analysed interventions, the intervention based on the most substantial and least heterogeneous body of literature appears to be structured parent involvement (parents given specific instructions on what to do with their infant during a painful procedure either verbally or through an information sheet). Based on low‐ to moderate‐certainty evidence, structured parent involvement (as studied) likely does not reduce pain behaviours in infants. However, an age effect may be suggested by the pattern of findings, such that structured parent involvement may reduce pain behaviours in older infants (i.e. 18 months) as opposed to younger (i.e. six to eight months). Of the 11 interventions described qualitatively, facilitated tucking, soothing sounds, multisensory bundles, video distraction, and the additive effects of a non‐pharmacological intervention onto topical anaesthetics may reduce pain behaviours in older infants, but the evidence base is of low to very low certainty. 

Adverse events

Of the 138 included studies, six reported adverse events (vomiting, abdominal distension, transient skin reactions, desaturations, infants becoming distressed, septicaemia), 70 reported that no adverse events occurred, and the remaining 62 did not report on adverse events at all. 

Overall completeness and applicability of evidence

Our goal was to update our 2015 review and assess the efficacy of all non‐pharmacological interventions, which had not been included in other Cochrane Reviews, for the management of acute pain and distress in infants and children up to three years of age. Given our evidence‐based understanding of the steep developmental trajectory that occurs during infancy and that this steep trajectory influences the efficacy of non‐pharmacological treatments, we took the additional step of disaggregating results for both age group (preterm born neonates, full‐term born neonates, and full‐term born older infants) and pain phase (reactivity, immediate regulation). While there was a considerable body of randomised controlled trials the evidence base is challenged by the lack of sufficient evidence for all the intervention strategies examined in this review.

The current review included people from both inpatient and outpatient settings and assessed a broad array of non‐pharmacological interventions, contributing to its overall completeness. Given that healthy and sick infants were included in the review, the findings are generalisable for healthy and typically developing infants as well as those hospitalised for various conditions. Results are only applicable and generalisable for the age and pain phase for which they were reported. 

Quality of the evidence

The level of confidence in the evidence in the review was generally low to very low. Few analyses had moderate certainty in the evidence, and none had high certainty. This finding behooves researchers and trialists to improve the quality of randomised controlled trials. Key reasons for the lack of reliable evidence for findings stems from small sample sizes within studies, the small number of studies within the same intervention, and the low quality of randomised controlled trials in the field (issues such as high risk of bias and imprecision). In light of these limitations, we used the best available evidence. Criteria for assessing certainty were based on recommendations from Cochrane, GRADE criteria, and the American Psychological Association (Chambless 2001).

Furthermore, as per Cochrane guidelines (Chapter 14, Higgins 2019), the current review examined the integrity of the interventions implemented (i.e. were both the intervention and care received by the control groups actually delivered as planned) in each included study as this can have important consequences for the study findings. One of the key reasons for poor intervention integrity related to the inadequate description of 'standard care'. As standard care can vary across clinical settings (i.e. some sites administer sucrose as part of their standard care while others do not), it is critical to provide this description to help contextualise the study results. Another reason for downgraded integrity ratings related to lack of reporting on infant positioning during painful procedures (i.e. if the infant lying in a supine, prone, or side‐lying position). This variable can impact infants' stress and pain levels during painful procedures and thus should be controlled for in RCTs (Wiley 2020). 

Of 138 studies included in this review, 103 were deemed to have a high risk of bias. We noted that over 50% of included studies had a high risk of bias in blinding of personnel, hence there is potential for performance bias. However, it should be noted that 69 of the 103 studies considered to have high risk of bias were only due to issues with blinding of personnel and outcome assessors. In other words, if blinding had not been assessed, these studies would not have been deemed high risk of bias. This type of evaluation is biased against RCTs examining non‐pharmacological pain management strategies as blinding of study personnel or outcome assessors is often not feasible in these trials due to the nature of the interventions (e.g. sucking on a pacifier; being swaddled or rocked; manipulating the environment, such as lights, sounds, temperature; distraction). Due to the non‐pharmacological nature of these interventions, it is a challenge that needs to be addressed in future review updates. In RCTs, there are five roles for which blinding should be considered: participants, personnel administering the intervention, personnel collecting data, outcome assessors, and data analysts. Moreover, a distinction should be made in future risk of bias tools between blinding to study hypotheses versus blinding to intervention. Although the nature of these interventions generally precludes personnel administering the intervention, collecting data, and outcome assessors to be blinded to the intervention, it is critical (and possible) that they be blinded to study hypotheses. It should be noted that, due to their developmental stage, the participants in these trials (infants) are blinded to both study hypotheses and intervention. However, if care‐givers are present during the study, the same distinction would apply whereas they should be blinded to study hypotheses even though they cannot usually also be blinded to their infants' assigned intervention. Finally, as it is feasible for data analysts to be blinded to both, this approach should be adopted in future trials. Overall, the use of an updated risk of bias tool to assess non‐pharmacological pain management trials, particularly with regard to the types of blinding, is warranted in future review updates.

Potential biases in the review process

Although we searched clinical trial registers for this review, we did not search other sources of non‐published data, such as dissertation databases, which may have introduced some bias. We sought non‐professional language speakers through list‐serves and Cochrane Task Exchange tool to fully extract and conduct risk of bias analyses for non‐English articles (six studies). If we were not able to find a translator, we placed the study in 'Awaiting classification' (two studies). In addition, this review included studies that assessed efficacy and adverse events, but not safety. We identified no other potential sources of bias.

Agreements and disagreements with other studies or reviews

Previous systematic reviews assessing pain management in children have focused on very specific forms of pain management strategies, such as music (Cepeda 2013), breast milk (Shah 2012), kangaroo care (Johnston 2017), and sucrose (Stevens 2016a), therefore we did not assess any of these pain management strategies for the purpose of this review. Another major distinction between this review and others in this research area is the division of pain response into pain reactivity and immediate pain regulation. A systematic review was done in order to create Clinical Practice Guidelines for Reducing Pain from Vaccine Injections in Children and Adults (Taddio 2015b), inclusive of included non‐pharmacological interventions, as well as the four above interventions. Past work has found that topical anaesthetics on acute distress had a SMD of –0.91 (95% CI –1.36 to –0.47) (Taddio 2015b), kangaroo care had a mean difference (MD) of ‐3.21 (95% CI ‐3.94 to ‐2.47; 5 studies, n = 267) (Johnston 2017), and sucrose had a weighted mean difference (WMD) of –1.70 (95% CI –2.13 to –1.26) (Stevens 2016a). The current review has shown effect sizes in similar ranges and was more complex in the analyses based on the consideration of both age and time of the 24 included strategies. Of the 18 analysed interventions, we found that 28% had no effect, 11% had a medium effect size, and 61% had a large effect size.

Authors' conclusions

Implications for practice.

Implications for care‐givers of infants in pain

We examined 24 different pain management interventions that utilised cognitive, behavioural, and contextual care‐giving strategies. We analysed 18 of these interventions separately for three different infant development stages (preterm born, full‐term neonate, older infants) and two different phases post‐acute painful procedure (reactivity and immediate regulation). One of the strengths of this review was that we divided the outcomes into three developmental stages within infancy and toddlerhood. Therefore, ill and/or premature babies are a group to themselves as there are notable differences related to reactivity and variability within this population. Specifically, ill and/or premature infants exhibit more subtle facial actions when signalling their pain‐related distress. To address this issue with pain assessment in preterms, we allowed for multidimensional pain measures in our review (behaviour and physiology) that are most commonly used clinically and have demonstrated good psychometric properties for preterm procedural pain (i.e. PIPP). In addition, some research has suggested divergence between behaviour, cardiac, and cortical measurements. Future work should be dedicated to integrating cortical measurements into clinical pain practice and understanding the circumstances contributing to the divergence of different types of pain measures.

Certainty in the findings was ascertained. While many strategies appeared to show some promise in reducing pain behaviours, none of the pain management strategies supported a rating of high certainty in their findings. Most evidence had very low to low certainty in the estimate of the findings. We highlight the interventions based on the most substantial body of literature and least heterogeneity below.

Preterm born neonates                       

Pain reactivity

Non‐nutritive sucking and facilitated tucking show the most promise for reducing pain reactivity in preterms, although this is based on an evidence base with very low certainty. 

Immediate pain regulation

Non‐nutritive sucking, swaddling, and facilitated tucking show the most promise for improving immediate pain regulation in preterms, although this is based on an evidence base with very low certainty. 

Full‐term born neonates                          

Pain reactivity and immediate pain regulation

Overall, non‐nutritive sucking showed the most promise for reducing pain behaviours in both pain phases for full‐term neonates, based on an evidence base with very low certainty.

Full‐term born older infants             

Pain reactivity and immediate pain regulation

Overall, structured parent involvement was the intervention most studied, and thus based on the most substantial body of evidence, in older infants. This evidence suggests that this intervention probably results in little to no difference in reducing pain behaviours in full‐term older infants. However, a pattern emerged such that this intervention appeared to be efficacious in older groups of infants within this category (i.e. 18 months), but not in younger infants (i.e. six to eight months). 

Due to the ethical considerations with the implementation of no‐treatment control groups, more research has emerged on additive trials. Based on the randomised controlled trials (RCTs) that were completed, we made decisions to coarsely cluster the additive effects of any non‐pharmacological strategies onto three types of 'base' strategies: sweet solutions, non‐nutritive sucking, and swaddling. While that allowed for a preliminary understanding of additive effects, the heterogeneity of interventions that were collapsed in this category limits the ability to discuss the additive effect of any non‐pharmacological intervention. As more RCTs examining the same combinations of interventions emerge in the field, a clearer picture of the enhanced efficacy of additive trials across age groups and pain phases will be obtained. We expect this to be an important trend in future trials, with a complete cessation of no‐treatment control pain management RCTs. With the above caveats in mind, it does appear that in preterm infants additive effects are evident when adding a non‐pharmacological intervention on top of a sweet solution or on top of non‐nutritive sucking.

Implications for policymakers

The data continue to strengthen the evidence base regarding unmitigated pain exposure with both young children who are hospitalised or not. This review provides a comprehensive inventory of strategies that should be considered for neonatal intensive care unit (NICU) or clinic practice guidelines on acutely painful procedures.

Implications for funders of the interventions

Funding agencies should examine proposals for new RCTs for non‐pharmacological interventions (e.g. touch‐related interventions, light reduction, and multisensory bundles) within the context of our specific methodological recommendations and the critical gaps that currently exist in the literature, which have been elucidated in this review.  Examples of key gaps are whether non‐nutritive sucking works for older infants, examining the role of rocking and holding in preterm infants, and why distraction and parent involvement are not effective in older infants/toddlers (contrary to what would be predicted by theory and practice in the acute pain context). Given the strength of the additive effect of sucrose, more research should be done to understand the impact of combining pharmacological and non‐pharmacological methods (e.g. sucrose and topical anaesthetics) on reducing infant pain.

• No‐treatment controls or inert controls (e.g. swallowing water) are now unethical in studies for infant pain and should not be used in any RCT methodology for pain.

• Future areas of non‐pharmacological approaches to managing infant pain research should therefore focus on increasing the quality of trials and either superiority trials (e.g. which of these two interventions are better), intervention trials requiring novel combinations, better implementation of existing interventions, and empowering parents with better soothing behaviours. For more detail read below in Implications for research.

Implications for research.

Based on the results of this review, important gaps in the existing literature on non‐pharmacological management of acute pain in young children are apparent. Amongst the gaps, the highest priority for researchers are suggested below.

1. None of the findings in this review are based on high confidence in the evidence. Greater attention must be paid to CONSORT reporting and the principles that subsume high‐quality trials. In addition to better reporting, there is a fundamental challenge relating to risk of bias scores in non‐pharmacological trials (e.g. how could one blind health professionals or behavioural coders to a child being given a soother or not). Future trials could be targeted (and explicit) in their attempts to mitigate these inherent biases by measuring factors that subsume health professional personnel bias (e.g. time to complete procedure, health professional affect, health professional time touching or verbalisations to infant during procedure) or outcome assessor bias (e.g. coders blinded to study hypotheses, having two coder teams (reliable with each other) code each arm). This would also require risk of bias tools being adapted to give 'credit' for credible attempts to mitigate these specific biases. To address these risk of bias nuances, future updates of this review will use Cochrane's RoB 2 tool. 

2. Unfortunately, the interventions that showed the most promise based on the most substantial body of literature were also based on uncertain or very uncertain evidence. This suggests that there should be a greater priority for high‐quality trials on these interventions: non‐nutritive sucking, swaddling, and facilitated tucking. Despite 138 trials in this review and many trials showing positive effects, few strategies could be recommended for pain management with moderate to high confidence in the evidence.

3. Several interventions (e.g. massage/touch‐based interventions across all ages, such as traditional massage or applying pressure to the procedure site; exposure to various soothing sounds in preterms, such as maternal voice or white noise) also emerged as potentially having a pain‐reducing effect; however, the evidence base for these interventions was very heterogenous, precluding us from drawing stronger conclusions about their effectiveness. As such, more high‐quality studies are needed on these interventions to build a more substantial evidence base. 

4. Structured care‐giver interventions (parent and non‐parent) need to be informed by attachment theory (Bowlby 1982). Over the first year of life, it has been argued that the care‐giver is the most important context for infants (Pillai Riddell 2009). Currently, studies that have attempted to formally structure parent behaviour have been limited and shown to be ineffective. It should be stressed that more work on better types of parent/non‐parent interventions, especially ones that capitalise on an infant’s primary developmental need for proximity to the parent (Bowlby 1982), is needed. Teaching care‐givers to better meet an infant’s attachment needs during times of pain may lead to more efficacious parent interventions. Parents are critical to understanding infant pain responses.

5. The lack of effects on distraction may be due to methodological challenges in the administration of distraction. Distraction is suggested to be best enacted after the peak distress has passed and the child's eyes are opened. Thus, while toy distraction was found to be not effective (with moderate confidence), it is recommended that greater attention be placed on examining the timing of the distraction.

6. Very few interventions were examined in healthy older infants. Given increasing publicity surrounding vaccination refusal and the importance of vaccination in a pandemic context, it is important to examine ways to reduce pain in this age group.

7. As aforementioned, a growing aspect of the review, based on changing trends in methodology and the ethical consideration of no‐treatment control trials, is the concept of 'additive' studies. As the field moves towards examining additive instead of no‐treatment control trials, future reviews should examine the additive effects of non‐pharmacological interventions on top of specific pharmacological interventions, such as topical anaesthetics (Bourdier 2019) or sucrose (Stevens 2016a). 

8. Observational research in a large longitudinal study that used latent class analysis techniques to analyse infant pain responses over time, substantiates the assertion that there are different subgroups of pain responders (Pillai Riddell 2013b), which render average pain estimates (i.e. the key outcome in most non‐pharmacological trials for acute pain) challenging to interpret due to the inherent variability within a group, particularly in the regulation phase. Using covariates such as attachment relationship, temperament or contextual factors is critical to move the field forward.

9. The current review solely focused on behavioural indicators of pain as these are most commonly used in clinical trials. In the future, cortical responses to pain are considered by researchers to be a leading potential indicator specific to pain‐related distress (Pillai Riddell 2016). However, current infant clinical pain assessment tools have not yet incorporated this dimension. Some research has demonstrated a link between behavioural and cortical pain indicators, however this association is highly influenced by context. The relationship between brain and behaviour have been found dependent on maternal positioning (Jones 2020), stress levels (Jones 2017), and post‐sucrose administration (Slater 2010). Therefore, infant pain assessment researchers must continue to improve the outcome measures used in pain management intervention trials and incorporating contextual influences.                          

In conclusion, reflecting on this update, the largest challenge with the literature is not the quantity of studies but rather the confidence in the work that has been done. To increase the confidence of findings, one approach for the next review update is to move forward without low‐quality trials. However, this would mean omitting most of the evidence and therefore is at odds with the mission of the review. Over the past 15 years, the team has striven to provide the most comprehensive literature synthesis possible for clinicians. Therefore, for future updates, we will continue to consider strategies to help strengthen our confidence in the evidence (i.e. the use of sub‐analyses and sensitivity analyses).

What's new

Date	Event	Description
14 June 2023	New search has been performed	This review has been updated to include the results of a new search on 8 October 2020. A further search was conducted in July 2022 and any additional potential included studies have been included in the Characteristics of studies awaiting classification table. These will be incorporated fully the next time this review is updated.
14 June 2023	New citation required and conclusions have changed	New intervention combinations were added to the review (sound addition, sound reduction, smell addition, light reduction, cool temperature addition, heat addition, multisensory bundle, additive effect of non‐pharmacological interventions on sugar solutions, additive effect of non‐pharmacological interventions on non‐nutritive sucking, additive effect of non‐pharmacological interventions on swaddling). Intervention combinations with fewer than two studies were omitted from the quantitative meta‐analysis, but described qualitatively. The search was updated on 8 October 2020 and all relevant studies assessed and analysed for inclusion in this update. We included a further 76 new studies making a total of 138 included studies (11,058 participants) for this update. We added 33 new studies to the Characteristics of studies awaiting classification table following a search from October 2022 to July 2022; these will be fully incorporated at the next review update. Some conclusions and confidence in findings differ from the 2015 review.

History

Protocol first published: Issue 4, 2006
Review first published: Issue 10, 2011

Date	Event	Description
4 March 2015	New citation required but conclusions have not changed	Kangaroo care section now omitted due to new 2013 Cochrane Review on topic. Eleven new interventions or intervention combinations added to review (structured non‐parent involvement; therapeutic touch; additive effect of sucking to co‐bedding + sucrose; additive effect of massage on sucrose + holding; additive effect of structured parent involvement on sucrose; additive effect of pacifier on sucrose; additive effect of facilitated tucking on sucking + sucrose, familiar odour, unfamiliar odour, co‐bedding and heel warming). Search updated on 4 March 2015. 21 new studies added (Badiee 2014; Basiri 2014; Campbell‐Yeo 2012; Chik 2012; Esfahani 2013; Harrington 2012; Herrington 2014; Hogan 2014; Johnston 2013; Jose 2012; Kristoffersen 2011; Liaw 2011; Liaw 2012; Liaw 2013; Park 2006; Mirzarahimi 2013; Sadathosseini 2013; Shaw 1982; Shu 2014; Sundaram 2013; Taddio 2015). 1595 new participants included in the review. GRADE evaluations now embedded in meta‐summary of findings (SoF) table and in individual SoF tables. All conclusions stayed same but confidence in the findings is elaborated. The application of GRADE criteria and the new types of interventions are compelling reasons to reread this update and especially the SoF tables.
4 March 2015	New search has been performed	This review has been updated to include the results of a new search, new risk of bias tables, and application of GRADE criteria; summary of findings tables have been added.
10 November 2008	Amended	Further RevMan 5 conversion changes.
28 July 2008	Amended	Converted to new review format.

Appendices

Appendix 1. CENTRAL (The Cochrane Library) 2022 update

#1MeSH descriptor: [Needles] this term only
#2needle*:ti,ab,kw (Word variations have been searched)
#3(blood sampl* or immuni* or inoculat* or vaccin* or inject* or "finger prick*" or finger‐prick or "heel prick*" or heel‐prick* or "heel lance*" or heel‐lance* or "heel puncture*" or heel‐puncture* or "heel stick" or suture* or (laceration* near/3 repair*)):ti,ab,kw (Word variations have been searched)
#4("lumbar puncture" or lumbar‐puncture* or "spinal tap*" or spinal‐tap*):ti,ab,kw (Word variations have been searched)
#5("bone marrow aspiration" or "bone marrow biops*"):ti,ab,kw (Word variations have been searched)
#6(intravenous or intra‐venous or venepuncture* or venipuncture* or venous cannulation* or (arterial blood gas* and cannul*)):ti,ab,kw (Word variations have been searched)
#7((catheter near/6 insert*) or catheter* or port‐a‐cath* or portacath):ti,ab,kw (Word variations have been searched)
#8("central line" near/6 (insert* or remov*)):ti,ab,kw (Word variations have been searched)
#9(central venous catheter* near/6 insert*):ti,ab,kw (Word variations have been searched)
#10(local analges* or local anaesthe* or local anesthe*):ti,ab,kw (Word variations have been searched)
#11((arterial puncture or artery) near/6 puncture*):ti,ab,kw (Word variations have been searched)
#12"arterial line*":ti,ab,kw (Word variations have been searched)
#13(thoracocentesis or paracentesis):ti,ab,kw (Word variations have been searched)
#14#1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11 or 12 or #13
#15MeSH descriptor: [Pain] explode all trees
#16MeSH descriptor: [Pain Measurement] this term only
#17MeSH descriptor: [Pain Threshold] this term only
#18pain*:ti,ab,kw (Word variations have been searched)
#19#15 or #16 or #17 or #18
#20#14 and #19
#21((vaccin* near/6 pain) or (cannul* near/6 pain) or (needle* near/6 pain*) or (needle* near/6 distress*) or (needle* near/6 discomfort) or (needle* near/6 fear*) or (needle* near/6 fright*) or (needle* near/6 anxious) or (needle* near/6 anxiet*) or (procedure* near/6 pain*) or (intervention* near/6 pain*) or (intervention* near/6 distress*) or (procedure near/6 distress*) or (procedure* near/6 discomfort*) or (procedure‐related near/6 pain)):ti,ab,kw (Word variations have been searched)
#22#20 or #21
#23MeSH descriptor: [Pain, Postoperative] this term only
#24((postoperative near/3 pain*) or (post‐operative near/3 pain*) or post‐operative‐pain):ti,ab,kw (Word variations have been searched)
#25((post‐surgical near/3 pain*) or ("post surgical" near/3 pain*) or (post‐surgery near/3 pain*) or (post adj surg* near/3 pain*)):ti,ab,kw (Word variations have been searched)
#26(post* near/4 pain*):ti,ab,kw (Word variations have been searched)
#27"pain relief after":ti,ab,kw (Word variations have been searched)
#28("pain following" near/3 surg*):ti,ab,kw (Word variations have been searched)
#29(posttreatment near/3 pain*):ti,ab,kw (Word variations have been searched)
#30(pain control after near/4 surg*):ti,ab,kw (Word variations have been searched)
#31((post surg* or post‐surg*) and (pain* or discomfort)):ti,ab,kw (Word variations have been searched)
#32((pain* near/3 after surg*) or (pain* near/3 after operat*)):ti,ab,kw (Word variations have been searched)
#33((pain* near/3 follow* operat*) or (pain* near/3 follow* surg*)):ti,ab,kw (Word variations have been searched)
#34#23 or #24 or #25 or #26 or #27 or #28 or #29 or #30 or #31 or #32 or #33
#35#22 or #34
#36MeSH descriptor: [Child, Preschool] this term only
#37MeSH descriptor: [Infant] explode all trees
#38(baby or babies or neonate* or newborn or child* or infant* or paediatric* or pediatric*):ti,ab,kw (Word variations have been searched)
#39#36 or #37 or #38
#40#35 and #39

Appendix 2. MEDLINE (OVID) search strategy 2022 update

1. Needles/
2. needle*.mp.
3. (blood sampl* or immuni* or inoculat* or vaccin* or inject* or "finger prick*" or finger‐prick or "heel prick*" or heel‐prick* or "heel lance*" or heel‐lance* or "heel puncture*" or heel‐puncture* or "heel stick" or suture* or (laceration* adj3 repair*)).mp.
4. ("lumbar puncture" or lumbar‐puncture* or "spinal tap*" or spinal‐tap*).mp.
5. ("bone marrow aspiration" or "bone marrow biops*").mp.
6. (intravenous or intra‐venous or venepuncture* or venipuncture* or venous cannulation* or (arterial blood gas* and cannul*)).mp.
7. ((catheter adj6 insert*) or catheter* or port‐a‐cath* or portacath).mp.
8. ("central line" adj6 (insert* or remov*)).mp.
9. (central venous catheter* adj6 insert*).mp.
10. (local analges* or local anaesthe* or local anesthe*).mp.
11. ((arterial puncture or artery) adj6 puncture*).mp.
12. "arterial line*".mp.
13. (thoracocentesis or paracentesis).mp.
14. or/1‐13
15. exp Pain/
16. Pain Measurement/
17. PAIN THRESHOLD/
18. pain*.mp.
19. or/15‐18
20. 14 and 19
21. ((vaccin* adj6 pain) or (cannul* adj6 pain) or (needle* adj6 pain*) or (needle* adj6 distress*) or (needle* adj6 discomfort) or (needle* adj6 fear*) or (needle* adj6 fright*) or (needle* adj6 anxious) or (needle* adj6 anxiet*) or (procedure* adj6 pain*) or (intervention* adj6 pain*) or (intervention* adj6 distress*) or (procedure adj6 distress*) or (procedure* adj6 discomfort*) or (procedure‐related adj6 pain)).mp.
22. or/20‐21
23. Pain, Postoperative/
24. ((postoperative adj3 pain*) or (post‐operative adj3 pain*) or post‐operative‐pain).mp.
25. ((post‐surgical adj3 pain*) or ("post surgical" adj3 pain*) or (post‐surgery adj3 pain*) or (post adj surg* adj3 pain*)).mp.
26. (post* adj4 pain*).mp.
27. "pain relief after".mp.
28. ("pain following" adj3 surg*).mp.
29. (posttreatment adj3 pain*).mp.
30. ("pain control after" adj4 surg*).mp.
31. ((post surg* or post‐surg*) and (pain* or discomfort)).mp.
32. ((pain* adj3 "after surg*") or (pain* adj3 "after operat*")).mp.
33. ((pain* adj3 "follow* operat*") or (pain* adj3 "follow* surg*")).mp.
34. or/23‐33
35. 22 or 34
36. Child, Preschool/
37. exp Infant/
38. (baby or babies or neonate* or newborn or child* or infant* or paediatric* or pediatric*).mp.
39. or/36‐38
40. 35 and 39
41. randomized controlled trial.pt.
42. controlled clinical trial.pt.
43. randomized.ab.
44. placebo.ab.
45. drug therapy.fs.
46. randomly.ab.
47. trial.ab.
48. or/41‐47
49. exp animals/ not humans.sh.
50. 48 not 49
51. 40 and 50

Appendix 3. EMBASE (OVID) search strategy 2022 update

1. Needles/
2. needle*.mp.
3. (blood sampl* or immuni* or inoculat* or vaccin* or inject* or "finger prick*" or finger‐prick or "heel prick*" or heel‐prick* or "heel lance*" or heel‐lance* or "heel puncture*" or heel‐puncture* or "heel stick" or suture* or (laceration* adj3 repair*)).mp.
4. ("lumbar puncture" or lumbar‐puncture* or "spinal tap*" or spinal‐tap*).mp.
5. ("bone marrow aspiration" or "bone marrow biops*").mp.
6. (intravenous or intra‐venous or venepuncture* or venipuncture* or venous cannulation* or (arterial blood gas* and cannul*)).mp.
7. ((catheter adj6 insert*) or catheter* or port‐a‐cath* or portacath).mp.
8. ("central line" adj6 (insert* or remov*)).mp.
9. (central venous catheter* adj6 insert*).mp.
10. (local analges* or local anaesthe* or local anesthe*).mp.
11. ((arterial puncture or artery) adj6 puncture*).mp.
12. "arterial line*".mp.
13. (thoracocentesis or paracentesis).mp.
14. or/1‐13
15. exp Pain/
16. Pain Measurement/
17. PAIN THRESHOLD/
18. pain*.mp.
19. or/15‐18
20. 14 and 19
21. ((vaccin* adj6 pain) or (cannul* adj6 pain) or (needle* adj6 pain*) or (needle* adj6 distress*) or (needle* adj6 discomfort) or (needle* adj6 fear*) or (needle* adj6 fright*) or (needle* adj6 anxious) or (needle* adj6 anxiet*) or (procedure* adj6 pain*) or (intervention* adj6 pain*) or (intervention* adj6 distress*) or (procedure adj6 distress*) or (procedure* adj6 discomfort*) or (procedure‐related adj6 pain)).mp.
22. or/20‐21
23. Pain, Postoperative/
24. ((postoperative adj3 pain*) or (post‐operative adj3 pain*) or post‐operative‐pain).mp.
25. ((post‐surgical adj3 pain*) or ("post surgical" adj3 pain*) or (post‐surgery adj3 pain*) or (post adj surg* adj3 pain*)).mp.
26. (post* adj4 pain*).mp.
27. "pain relief after".mp.
28. ("pain following" adj3 surg*).mp.
29. (posttreatment adj3 pain*).mp.
30. ("pain control after" adj4 surg*).mp.
31. ((post surg* or post‐surg*) and (pain* or discomfort)).mp.
32. ((pain* adj3 "after surg*") or (pain* adj3 "after operat*")).mp.
33. ((pain* adj3 "follow* operat*") or (pain* adj3 "follow* surg*")).mp.
34. or/23‐33
35. 22 or 34
36. Child, Preschool/
37. exp Infant/
38. (baby or babies or neonate* or newborn or child* or infant* or paediatric* or pediatric*).mp.
39. or/36‐38
40. 35 and 39
41. random$.tw.
42. factorial$.tw.
43. crossover$.tw.
44. cross over$.tw.
45. cross‐over$.tw.
46. placebo$.tw.
47. (doubl$ adj blind$).tw.
48. (singl$ adj blind$).tw.
49. assign$.tw.
50. allocat$.tw.
51. volunteer$.tw.
52. Crossover Procedure/
53. double‐blind procedure.tw.
54. Randomized Controlled Trial/
55. Single Blind Procedure/
56. or/41‐55
57. (animal/ or nonhuman/) not human/
58. 56 not 57
59. 40 and 58

Appendix 4. PsycINFO (OVID) search strategy 2022 update

1. needle*.mp.
2. (blood sampl* or immuni* or inoculat* or vaccin* or inject* or "finger prick*" or finger‐prick or "heel prick*" or heel‐prick* or "heel lance*" or heel‐lance* or "heel puncture*" or heel‐puncture* or "heel stick" or suture* or (laceration* adj3 repair*)).mp.
3. ("lumbar puncture" or lumbar‐puncture* or "spinal tap*" or spinal‐tap*).mp.
4. ("bone marrow aspiration" or "bone marrow biops*").mp.
5. (intravenous or intra‐venous or venepuncture* or venipuncture* or venous cannulation* or (arterial blood gas* and cannul*)).mp.
6. ((catheter adj6 insert*) or catheter* or port‐a‐cath* or portacath).mp.
7. ("central line" adj6 (insert* or remov*)).mp.
8. (central venous catheter* adj6 insert*).mp.
9. (local analges* or local anaesthe* or local anesthe*).mp.
10. ((arterial puncture or artery) adj6 puncture*).mp.
11. "arterial line*".mp.
12. (thoracocentesis or paracentesis).mp.
13. or/1‐12
14. exp Pain/
15. Pain Measurement/
16. PAIN THRESHOLD/
17. pain*.mp.
18. or/14‐17
19. 13 and 18
20. ((vaccin* adj6 pain) or (cannul* adj6 pain) or (needle* adj6 pain*) or (needle* adj6 distress*) or (needle* adj6 discomfort) or (needle* adj6 fear*) or (needle* adj6 fright*) or (needle* adj6 anxious) or (needle* adj6 anxiet*) or (procedure* adj6 pain*) or (intervention* adj6 pain*) or (intervention* adj6 distress*) or (procedure adj6 distress*) or (procedure* adj6 discomfort*) or (procedure‐related adj6 pain)).mp.
21. or/19‐20
22. ((postoperative adj3 pain*) or (post‐operative adj3 pain*) or post‐operative‐pain).mp.
23. ((post‐surgical adj3 pain*) or ("post surgical" adj3 pain*) or (post‐surgery adj3 pain*) or (post adj surg* adj3 pain*)).mp.
24. (post* adj4 pain*).mp.
25. "pain relief after".mp.
26. ("pain following" adj3 surg*).mp.
27. (posttreatment adj3 pain*).mp.
28. ("pain control after" adj4 surg*).mp.
29. ((post surg* or post‐surg*) and (pain* or discomfort)).mp.
30. ((pain* adj3 "after surg*") or (pain* adj3 "after operat*")).mp.
31. ((pain* adj3 "follow* operat*") or (pain* adj3 "follow* surg*")).mp.
32. or/22‐31
33. 21 or 32
34. (baby or babies or neonate* or newborn or child* or infant* or paediatric* or pediatric*).mp.
35. 33 and 34
36. clinical trials/
37. (randomis* or randomiz*).tw.
38. (random$ adj3 (allocat$ or assign$)).tw.
39. ((clinic$ or control$) adj trial$).tw.
40. ((singl$ or doubl$ or trebl$ or tripl$) adj3 (blind$ or mask$)).tw.
41. (crossover$ or "cross over$").tw.
42. random sampling/
43. Experiment Controls/
44. Placebo/
45. placebo$.tw.
46. exp program evaluation/
47. treatment effectiveness evaluation/
48. ((effectiveness or evaluat$) adj3 (stud$ or research$)).tw.
49. or/36‐48
50. 35 and 49

Appendix 5. CINAHL (EBSCO) search strategy 2022 update

S50 S40 AND S49
S49 S41 OR S42 OR S43 OR S44 OR S45 OR S46 OR S47 OR S48
S48 (allocat* random*)
S47 (MH "Quantitative Studies")
S46 (MH "Placebos")
S45 placebo*
S44 (random* allocat*)
S43 (MH "Random Assignment")
S42 (Randomi?ed control* trial*)
S41 (singl* blind* ) or (doubl* blind* ) or (tripl* blind* ) or (trebl* blind* ) or (trebl* mask* ) or (tripl* mask* ) or (doubl* mask* ) or (singl* mask* )
S40 S35 AND S39
S39 S36 OR S37 OR S38
S38 (baby or babies or neonate* or newborn or child* or infant* or paediatric* or pediatric*)
S37 (MH "Infant+")
S36 (MH "Child, Preschool")
S35 S22 OR S34
S34 S23 OR S24 OR S25 OR S26 OR S27 OR S28 OR S29 OR S30 OR S31 OR S32 OR S33
S33 ((pain* N3 "follow* operat*") or (pain* N3 "follow* surg*"))
S32 ((pain* N3 "after surg*") or (pain* N3 "after operat*"))
S31 ((post surg* or post‐surg*) and (pain* or discomfort))
S30 ("pain control after" N4 surg*)
S29 (posttreatment N3 pain*)
S28 ("pain following" N3 surg*)
S27 "pain relief after"
S26 (post* N4 pain*)
S25 ((post‐surgical N3 pain*) or ("post surgical" N3 pain*) or (post‐surgery N3 pain*) or (post adj surg* N3 pain*))
S24 ((postoperative N3 pain*) or (post‐operative N3 pain*) or post‐operative‐pain)
S23 (MH "Postoperative Pain")
S22 S20 OR S21
S21 ((vaccin* N6 pain) or (cannul* N6 pain) or (needle* N6 pain*) or (needle* N6 distress*) or (needle* N6 discomfort) or (needle* N6 fear*) or (needle* N6 fright*) or (needle* N/anxious) or (needle* N6 anxiet*) or (procedure* N6 pain*) or (intervention* N6 pain*) or (intervention* N6 distress*) or (procedure N6 distress*) or (procedure* N6 discomfort*) or (procedure‐related N6 pain))
S20 S14 AND S19
S19 S15 OR S16 OR S17 OR S18
S18 pain*
S17 (MH "Pain Threshold")
S16 (MH "Pain Measurement")
S15 (MH "Pain+")
S14 S1 OR S2 OR S3 OR S4 OR S5 OR S6 OR S7 OR S8 OR S9 OR S10 OR S11 OR S12 OR S13
S13 (thoracocentesis or paracentesis)
S12 "arterial line*"
S11 ((arterial puncture or artery) N6 puncture*)
S10 (local analges* or local anaesthe* or local anesthe*)
S9 (central venous catheter* N6 insert*)
S8 ("central line" N6 (insert* or remov*))
S7 ((catheter N6 insert*) or catheter* or port‐a‐cath* or portacath)
S6 (intravenous or intra‐venous or venepuncture* or venipuncture* or venous cannulation* or (arterial blood gas* and cannul*))
S5 ("bone marrow aspiration" or "bone marrow biops*")
S4 ("lumbar puncture" or lumbar‐puncture* or "spinal tap*" or spinal‐tap*)
S3 (blood sampl* or immuni* or inoculat* or vaccin* or inject* or "finger prick*" or finger‐prick or "heel prick*" or heel‐prick* or "heel lance*" or heel‐lance* or "heel puncture*" or heel‐puncture* or "heel stick" or suture* or (laceration* N3 repair*))
S2 needle*
S1 (MH "Needles")

Data and analyses

Comparison 1. Non‐nutritive sucking X preterm neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Pain reactivity	7	597	Std. Mean Difference (IV, Random, 95% CI)	‐0.57 [‐1.03, ‐0.11]
1.2 Immediate pain regulation	6	379	Std. Mean Difference (IV, Random, 95% CI)	‐0.61 [‐0.95, ‐0.27]

Comparison 2. Non‐nutritive sucking X full‐term neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Pain reactivity	8	545	Std. Mean Difference (IV, Random, 95% CI)	‐1.13 [‐1.57, ‐0.68]
2.2 Immediate pain regulation	9	536	Std. Mean Difference (IV, Random, 95% CI)	‐1.49 [‐2.20, ‐0.78]

Comparison 3. Swallowing water X preterm neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Pain reactivity	2	120	Std. Mean Difference (IV, Random, 95% CI)	1.10 [‐1.56, 3.75]

Comparison 4. Rocking/holding X full‐term neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Pain reactivity	3	221	Std. Mean Difference (IV, Random, 95% CI)	‐0.09 [‐0.61, 0.43]
4.2 Immediate pain regulation	3	171	Std. Mean Difference (IV, Random, 95% CI)	‐0.84 [‐1.15, ‐0.53]

Comparison 5. Swaddling X preterm neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 Pain reactivity	5	468	Std. Mean Difference (IV, Random, 95% CI)	‐0.60 [‐1.23, 0.04]
5.2 Immediate pain regulation	4	204	Std. Mean Difference (IV, Random, 95% CI)	‐1.21 [‐2.05, ‐0.38]

Comparison 6. Swaddling X full‐term neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
6.1 Pain reactivity	4	263	Std. Mean Difference (IV, Random, 95% CI)	‐0.89 [‐1.24, ‐0.55]

Comparison 7. Facilitated tucking X preterm neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
7.1 Pain reactivity	12	733	Std. Mean Difference (IV, Random, 95% CI)	‐1.01 [‐1.44, ‐0.58]
7.2 Immediate pain regulation	10	557	Std. Mean Difference (IV, Random, 95% CI)	‐0.59 [‐0.92, ‐0.26]

Comparison 8. Touch/massage related X preterm neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
8.1 Pain reactivity	3	230	Std. Mean Difference (IV, Random, 95% CI)	‐0.92 [‐1.60, ‐0.25]
8.2 Immediate pain regulation	2	112	Std. Mean Difference (IV, Random, 95% CI)	‐1.24 [‐1.85, ‐0.63]

Comparison 9. Touch/massage related X full‐term neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
9.1 Pain reactivity	6	520	Std. Mean Difference (IV, Random, 95% CI)	‐0.81 [‐1.22, ‐0.40]
9.2 Immediate pain regulation	3	295	Std. Mean Difference (IV, Random, 95% CI)	‐0.92 [‐1.57, ‐0.28]

Comparison 10. Touch/massage related X older infant.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
10.1 Pain reactivity	5	368	Std. Mean Difference (IV, Random, 95% CI)	‐1.28 [‐1.91, ‐0.65]
10.2 Immediate pain regulation	3	244	Std. Mean Difference (IV, Random, 95% CI)	‐1.45 [‐2.04, ‐0.86]

Comparison 11. Toy distraction x older infant.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
11.1 Pain reactivity	4	293	Std. Mean Difference (IV, Random, 95% CI)	‐0.15 [‐0.39, 0.08]
11.2 Immediate pain regulation	1	99	Std. Mean Difference (IV, Random, 95% CI)	‐0.08 [‐0.50, 0.34]

Comparison 12. Structured parent involvement X older infant.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
12.1 Pain reactivity	6	653	Std. Mean Difference (IV, Random, 95% CI)	‐0.18 [‐0.40, 0.03]
12.2 Immediate pain regulation	4	728	Std. Mean Difference (IV, Random, 95% CI)	‐0.09 [‐0.40, 0.21]

Comparison 13. Sound reduction X preterm neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
13.1 Pain reactivity	2	144	Std. Mean Difference (IV, Random, 95% CI)	‐0.64 [‐2.06, 0.78]

Comparison 14. Sound addition X preterm neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
14.1 Pain reactivity	7	362	Std. Mean Difference (IV, Random, 95% CI)	‐0.89 [‐1.43, ‐0.34]
14.2 Immediate pain regulation	4	360	Std. Mean Difference (IV, Random, 95% CI)	‐0.92 [‐2.20, 0.36]

Comparison 15. Smell addition X preterm neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
15.1 Pain reactivity	5	383	Std. Mean Difference (IV, Random, 95% CI)	‐0.41 [‐0.72, ‐0.10]
15.2 Immediate pain regulation	5	383	Std. Mean Difference (IV, Random, 95% CI)	‐0.57 [‐0.88, ‐0.27]

Comparison 16. Smell addition X full‐term neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
16.1 Pain reactivity	2	182	Std. Mean Difference (IV, Random, 95% CI)	‐0.77 [‐1.09, ‐0.44]
16.2 Immediate pain regulation	2	237	Std. Mean Difference (IV, Random, 95% CI)	‐0.41 [‐0.80, ‐0.01]

Comparison 17. Light reduction x preterm neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
17.1 Pain reactivity	2	125	Std. Mean Difference (IV, Random, 95% CI)	‐0.71 [‐1.08, ‐0.34]
17.2 Immediate pain regulation	2	125	Std. Mean Difference (IV, Random, 95% CI)	‐1.16 [‐1.53, ‐0.78]

Comparison 18. Heat addition X full‐term neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
18.1 Pain reactivity	3	232	Std. Mean Difference (IV, Random, 95% CI)	‐0.12 [‐0.42, 0.17]

Comparison 19. Cold addition X full‐term neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
19.1 Pain reactivity	2	142	Std. Mean Difference (IV, Random, 95% CI)	‐0.85 [‐1.48, ‐0.23]

Comparison 20. Multisensory bundle X preterm neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
20.1 Pain reactivity	5	280	Std. Mean Difference (IV, Random, 95% CI)	‐2.70 [‐4.16, ‐1.23]
20.2 Immediate pain regulation	4	228	Std. Mean Difference (IV, Random, 95% CI)	‐1.49 [‐2.92, ‐0.06]

Comparison 21. Multisensory bundle X full‐term neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
21.1 Pain reactivity	2	227	Std. Mean Difference (IV, Random, 95% CI)	0.00 [‐0.29, 0.30]

Comparison 22. Additive non‐pharmacological intervention on sweet solution X preterm neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
22.1 Pain reactivity	7	456	Std. Mean Difference (IV, Random, 95% CI)	‐1.10 [‐1.96, ‐0.25]
22.2 Immediate pain regulation	4	411	Std. Mean Difference (IV, Random, 95% CI)	‐0.90 [‐2.00, 0.20]

Comparison 23. Additive non‐pharmacological intervention on sweet solution X full‐term neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
23.1 Pain reactivity	4	263	Std. Mean Difference (IV, Random, 95% CI)	‐1.26 [‐2.45, ‐0.07]
23.2 Immediate pain regulation	5	875	Std. Mean Difference (IV, Random, 95% CI)	‐0.53 [‐0.93, ‐0.13]

Comparison 24. Additive non‐pharmacological intervention on sweet solution X older infant.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
24.1 Pain reactivity	2	241	Std. Mean Difference (IV, Random, 95% CI)	‐0.14 [‐0.39, 0.10]

Comparison 25. Additive non‐pharmacological intervention on non‐nutritive sucking X preterm neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
25.1 Pain reactivity	5	301	Std. Mean Difference (IV, Random, 95% CI)	‐1.27 [‐2.21, ‐0.33]
25.2 Immediate pain regulation	3	209	Std. Mean Difference (IV, Random, 95% CI)	‐1.32 [‐2.59, ‐0.05]

Comparison 26. Additive non‐pharmacological intervention on swaddling X full‐term neonate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
26.1 Pain reactivity	2	180	Std. Mean Difference (IV, Random, 95% CI)	‐0.98 [‐2.40, 0.44]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Abbasoglu 2015.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 32 Setting: hospital Diagnostic criteria: stable preterm infants Gestational age: Mean: 32.2 weeks SD: not reported Sex: Males: 12 Females: 20 Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: routine care Total number of participants randomised to each of the groups: Intervention: 16 Control: 16 Specific intervention: acupressure Intervention details: acupressure applied for 3 minutes at points BL60 and K3; points were kneaded and firmly stimulated in a gentle manner Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: acupressure at points BL60 and K3 was associated with shorter procedural times and shorter crying duration in preterm infants Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Authors do not sufficiently describe the method used to generate the randomisation.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	Unclear risk	There is no mention of blinding procedures.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	Unclear whether the video recording captures the foot, which would reveal the intervention.
4) Attrition bias (incomplete outcome data)	Low risk	The number of enrolled participants in each group and the number reported in outcome tables match.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, if one or more domain is "unclear", the overall judgement is unclear.

Akcan 2016.

Study characteristics
Methods	Study design: between‐group Total study duration: 3 months
Participants	Total number: 102 Setting: hospital Diagnostic criteria: full‐term healthy neonates Postnatal age: Mean: 2 days SD: not reported Sex: Males: 48 Females: 54 Country: Turkey
Interventions	Total number of intervention groups: 3 Control group description: distilled water odour Total number of participants randomised to each of the groups: Lavender odour: 27 Breast milk odour: 24 Amniotic fluid odour: 26 Control (distilled water odour): 25 Specific intervention #1: lavender odour Intervention details #1: odour sample was held at 10 cm distance from infant's nose from 5 minutes before procedure until 5 minutes after Integrity of intervention #1: good Specific intervention #2: breast milk odour Intervention details #2: odour sample was held at 10 cm distance from infant's nose from 5 minutes before procedure until 5 minutes after Integrity of intervention #2: good Specific intervention #3: amniotic fluid odour Intervention details #3: odour sample was held at 10 cm distance from infant's nose from 5 minutes before procedure until 5 minutes after Integrity of intervention #3: good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: first 30 seconds post‐procedure and 1 minute post‐procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: smell of lavender and breast milk reduced pain during the invasive procedure more then amniotic fluid and the control group Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	Envelopes used were opaque, but there is no mention of whether they were sealed or sequentially numbered. There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	Unclear risk	It is not clear whether the nurse could smell the different interventions.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	It is not clear whether the outcome assessor could smell the interventions.
4) Attrition bias (incomplete outcome data)	Low risk	No flow chart provided but the sample sizes reported in the Methods section match the Ns provided in the NIPS table so no reason to believe there were missing outcome data.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, a rating of "unclear" in one or more domain leads to an overall rating of unclear.

Alemdar 2017.

Study characteristics
Methods	Study design: between‐group Total study duration: 10 months
Participants	Total number: 94 Setting: hospital Diagnostic criteria: stable preterm infants Gestational age: Mean: 32.6 (intrauterine sounds), 32.2 (covered eyes), 32.8 (control), 32.6 (overall) SD: 3.1 (intrauterine sounds), 3.4 (covered eyes), 3.5 (control) Sex: Males: 59 Females: 33  Country: Turkey
Interventions	Total number of intervention groups: 2 Control group description: did not receive any intervention Total number of participants randomised to each of the groups: Intrauterine sounds: 32 Covered eyes: 32 Control: 30 Specific intervention #1: intrauterine sounds Intervention details #1: listened to intrauterine sounds (relaxing music similar to a real heartbeat converted to resemble intrauterine sounds) for 15 minutes before and up to 15 minutes after venipuncture Integrity of intervention #1: good Specific intervention #2: covered eyes Intervention details #2: eyes were covered by eye patches for 15 minutes before and up to 15 minutes after venipuncture Integrity of intervention #2: good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: first 30 seconds post‐procedure and 1 minute post‐procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: the practice of covering eyes during venipuncture positively affected infants' pain scores after venipuncture Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	It would be impossible to blind personnel to the intervention due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	Could not record videos without capturing infants' faces and seeing whose eyes were closed.
4) Attrition bias (incomplete outcome data)	Low risk	Missing outcome data appear to be even across groups (with the exception of the control group).
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain is designated as "high risk" then there is an overall rating of high risk.

Alemdar 2017a.

Study characteristics
Methods	Study design: between‐group Total study duration: 14 months
Participants	Total number: 85 Setting: hospital Diagnostic criteria: stable preterm infants Gestational age: Mean: 34 (amniotic fluid), 32.1 (breast milk odour), 33.4 (control), 33.1 (mother odour) 33.1 (overall) SD: not reported Sex: Males: 16 (amniotic fluid), 15 (breast milk odour), 12 (mother odour), 12 (control), 55 (overall) Females: 5 (amniotic fluid), 9 (breast milk odour), 8 (mother odour), 8 (control), 30 (overall) Country: Turkey
Interventions	Total number of intervention groups: 3 Control group description: no intervention was given Total number of participants randomised to each of the groups: Amniotic fluid: 21 Breast milk odour: 22 Mother odour: 20 Control: 22 Specific intervention #1: amniotic fluid Intervention details #1: odour was placed 5 cm away from infants' noses by pouring it onto sterilised sponge made of gauze 15 minutes before the heel stick and up to 15 minutes after Integrity of intervention #1: good Specific intervention #2: breast milk odour Intervention details #2: odour was placed 5 cm away from infants' noses by pouring it onto sterilised sponge made of gauze 15 minutes before the heel stick and up to 15 minutes after Integrity of intervention #2: good Specific intervention #3: mother odour Intervention details #3: mother asked to hold doll on chest for a night to absorb odour and doll was position 5 cm away from infants' noses for 15 minutes before heel stick and up to 15 minutes after Integrity of intervention #3: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure and 1 minute post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: amniotic fluid, mother's milk, and mother's odour were not effective for preterm infants during painful procedures Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	It would be impossible to blind personnel to the intervention due to nature of intervention (soothing smell versus no smell).
3b) Detection bias (blinding of outcome assessors)	High risk	Although a NICU nurse not involved in the research scored the videos and then compared their answers to the PI's, the videos were still not blinded and the PI's scores were the ones added after all.
4) Attrition bias (incomplete outcome data)	Unclear risk	Attrition data provided in Fig. 1 is for a different study, not the current one (different sample sizes and interventions are reported); therefore, no actual explanation of attrition and missing data provided for the current sample, and it was unclear how missing participants or data were accounted for.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain is designated as "high risk" then there is an overall rating of high risk.

Alemdar 2018.

Study characteristics
Methods	Study design: between‐group Total study duration: not mentioned
Participants	Total number: 123 Setting: hospital Diagnostic criteria: preterm Gestational age: Mean: 30.1 (recorded maternal voice), 30.3 (breast milk odour), 30.2 (covered incubator), 30.3 (control), 30.2 (overall) SD: 0.6 (recorded maternal voice), 0.7 (breast milk odour), 0.7 (covered incubator), 0.5 (control), 0.6 (overall) Sex: Males: 62 Females: 61 Country: Turkey
Interventions	Total number of intervention groups: 3 Control group description: no pharmacological or non‐pharmacological interventions were applied Total number of participants randomised to each of the groups: Recorded maternal voice: 30 Breast milk odour: 30 Covered incubator: 31 Control: 32 Specific intervention #1: recorded maternal voice Intervention details #1: the mother of each preterm infant was encouraged to express her thoughts and feelings and to say what she wanted to say to her baby. As she spoke, her words were recorded by using a voice recorder. Then, this recorder was set to 45 dB and played for the infant inside an incubator for 15 minutes before the procedure. The recording was then played during and 15 minutes after the procedure Integrity of intervention #1: good Specific intervention #2: breast milk odour Intervention details #2: breast milk was taken from the mothers with the help of a milking machine when the mother was stable. A total of 5 cm3 of the milk sample received was poured in a sterile sponge 15 minutes before the procedure and placed 5 cm away from the infant. The infant continued to smell the milk during and 15 minutes after the procedure. Integrity of intervention #2: good Specific intervention #3: covered incubator Intervention details #3: incubators were covered with a cover that was made with a special thick white fabric, and the front side was left open. The incubators were covered before, during, and after the procedure. Integrity of intervention #3: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure and 1 minute post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors: breast milk odour, recorded maternal voice, and incubator cover in preterm infants are recommended as simple, safe, and supportive stimuli that facilitate positive effects during painful procedures Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Same nurse performed all peripheral cannulations and there's no mention of them being blinded. Also, it would be hard to blind since the conditions are so different: in one, a sponge is held next to baby's face, in the second a recoding is being played in the incubator (which the nurse would be able to hear), and for the third the incubator was covered.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors unable to be blinded.
4) Attrition bias (incomplete outcome data)	Low risk	Included flow chart explaining reasons for attrition and final #s for each group. Reason for attrition similar across groups and similar #s excluded from each group.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, there is at least one high rating, and therefore the overall summary judgement is high.

Alemdar 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: 16 months
Participants	Total number: 61 Setting: hospital Diagnostic criteria: preterm Gestational age: Mean: 31.3 (smelling amniotic fluid), 33.9 (control), overall mean = 32.6 SD: 2.6 (smelling amniotic fluid), 3.2 (control)   Sex: Males: 20 (smelling amniotic fluid), 18 (control), total = 38 Females: 10 (smelling amniotic fluid), 13 (control), total = 23 Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: peripheral cannulation was performed by clinical routine for premature babies in the control group. NOTE: The environment outside the incubator had similar properties for all infants: room temperature of 24 to 25 degree C (min‐max), ambient light of the 92 to 115 lx (min‐max), and the sound quantities of 48 to 82 db (min‐max) Total number of participants randomised to each of the groups: Intervention: 30 Control: 31 Specific intervention: smelling amniotic fluid Intervention details: smelled their own mothers' amniotic fluid: 8 cc to 10 cc of amniotic fluid was taken by the doctor into a 20 mL injector during caesarean section and kept at 8 degrees C in the refrigerator until cannulation administration. In premature infants hospitalised at the NICU, amniotic fluid was fixed 5 cm away from the infants' noses by pouring it on a big piece of sterilised gauze sponge 15 minutes before peripheral cannulation. The infants were made to smell amniotic fluid during the procedure and up to 15 minutes after the procedure. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: before, during, and after peripheral cannulation Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: the intervention of smelling amniotic fluid is a practice that can be used for reducing pain and stress of preterm infants during peripheral cannulation Study funding source: The Scientific Research Projects (SRP) of Management Unit of Giresun University Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel unable to be blinded due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors unable to be blinded due to nature of intervention.
4) Attrition bias (incomplete outcome data)	High risk	It appears that infants were excluded from analysis if they did not have sufficient saliva for cortisol analysis, even though they could have been retained in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Alinejad‐Naeini 2014.

Study characteristics
Methods	Study design: cross‐over Total study duration: 5 months
Participants	Total number: 34 Setting: hospital Diagnostic criteria: some full‐term neonates, some preterm neonates Gestational age: 55.9% < 32 weeks, 20.6% 32 to 36 weeks, 23.5% > 36 weeks Mean: not mentioned SD: not mentioned Sex: Males: 18 Females: 16 Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: in cases when the neonates were prescribed to receive suctioning without intervention, the research assistant performed the suction and the researcher only recorded the study variables. Total number of participants randomised to each of the groups: 34 Specific intervention: facilitated tucking Intervention details: the neonate was placed on his/her side, his/her back was gently bent, the legs were getting up and flexion at an angle greater than 90%, the shoulders were also constricted up to 90 degrees and the researcher's hands were placed over the head close to the mouth or on the neonate's face. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors: the facilitated tucking position can be used as a safe non‐pharmacological method for procedural pain management Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Authors do not sufficiently describe the method used to generate the randomisation.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	The research assistant who performed the suctioning was aware of the condition. Further, due to the covert nature of the intervention personnel would need to be aware of study condition.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were blinded to study condition.
4) Attrition bias (incomplete outcome data)	Unclear risk	Chart depicting flow of study does not report any sample sizes so hard to assess whether there was any attrition or if the data for all 34 infants were included in the analysis. Although the frequencies reported in the demographics table add up to 34, no sample sizes are given for the t test comparing PIPP means (no dfs are mentioned either). It was unclear how missing participants or data were accounted for.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall assessment is high risk.

Allen 1996.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 64 Setting: hospital ambulatory paediatric clinic Diagnostic criteria: healthy neonates and infants Age: 2 weeks old, 2 months old, 4 months old, 6 months old, 9 months old, 15 months old, 18 months old Sex: not reported Country: United States
Interventions	Total number of intervention groups: 1 Control group description: no intervention Total number of participants randomised to each group: Control = 32 Treatment = 32 Specific intervention: sterile water Intervention details: infants orally received 2 mL of sterile water 2 minutes before needle Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured Pain vocalisation (cry) Upper limit: 100% Lower limit: 0% Time points: video taping began 30 seconds before injection and continued 3 minutes afterward. The proportion of crying was recorded at 15‐second intervals High score = more pain (longer cry) Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: water administered prior to needle lowers pain response only when one needle is given in infants aged 2 weeks, 9 months, and 18 months Study funding source: supported in part by grant MCJ 319152 from the Maternal and Child Health Bureau, Health Resources Services Administration, Rockville, Md, and by grant 90 DD 032402 from the Administration on Developmental Disabilities, Washington, DC Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported, but this did not involve an independent person.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Unclear risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
4) Attrition bias (incomplete outcome data)	High risk	There is no documented evidence or insufficient evidence reported of how attrition was dealt with.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	High risk	Either equivalence of groups is not reported or there is evidence of non‐equivalence.
Total	High risk	As per Sterne 2019, total score on Cochrane standard: low risk (1/7), high risk (3/7), unclear risk (3/7).

Angeles 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: 50 months
Participants	Total number: 99 Setting: hospital Diagnostic criteria: premature Age: Mean: 31.3 (30% dextrose), 31.4 (30% dextrose and facilitated tucking) SD: 2.4 (30% dextrose), 2.1 (30% dextrose and facilitated tucking)   Sex: Males: 30 (30% dextrose), 28 (30% dextrose and facilitated tucking) Females: 17 (30% dextrose), 24 (30% dextrose and facilitated tucking) Country: United States
Interventions	Total number of intervention groups: 1 Control group description: oral dextrose was prepared by the clinical pharmacist and was administered over 60 seconds via syringe to the anterior tongue 2 minutes prior to the heel lance followed by giving the participant a pacifier Total number of participants randomised to each of the groups: 30% dextrose = 47, 30% dextrose and facilitated tucking = 52 Specific intervention: 30% dextrose and facilitated tucking Intervention details: not described in paper Integrity of intervention: poor
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP‐R Time points: pain scores were measured at baseline (0 minutes) and at the time of the heel lance Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: a single dose of 30% oral dextrose, given before a clinically required heel lance, decreased signs of pain without increasing ATP utilisation and oxidative stress in premature neonates Study funding source: this work was supported by NIH grant RO1 NR011209 (Angeles) Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	No information on blinding personnel provided and due to nature of intervention it is assumed that personnel were not blinded.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	It is unclear if outcomes assessors were blinded.
4) Attrition bias (incomplete outcome data)	Low risk	Showed attrition reasons that appear balanced across groups.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Apaydin Cirik 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: 12 months
Participants	Total number: 95 Setting: hospital Diagnostic criteria: preterm and being fed via OGT Age (days): Mean: 6.1 SD: 5.5 Sex: Males: 55.6% Females: 44.4% Country: Turkey
Interventions	Total number of intervention groups: 2 Control group description: in the routine care group, while the senior clinical neonatal nurse performed the insertion procedure, physiological measurements of the highest value of heart rate and the lowest value of oxygen saturation were recorded by one researcher, 1 minute before the procedure, during the process, and 1 and 2 minutes after, using an individual monitor Total number of participants randomised to each of the groups: Swaddling: 30 Facilitated tucking: 32 Control: 33 Specific intervention #1: swaddling Intervention details: the infants were placed in a supine position on a 90 cm 2 soft fabric bandage for swaddling. The upper side of the bandage was aligned with the shoulder, and the arms were put in a flexion position. The top of the body was completely covered with the bandage, and the head left able to move freely. The swaddling procedure took about 1 minute. The infants remained in the swaddling during the procedure and for 5 minutes after. Integrity of intervention #1: good Specific intervention #2: facilitated tucking Intervention details: given in the supine position by holding the arm and legs of the infant to the body of the researcher. Initiated 3 minutes before tube insertion in order for the infant to feel the fetal position and to cope with the painful procedure. The infant was kept in the fetal position during the procedure and for 5 minutes after. Integrity of intervention #1: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: 10 minutes before, 2 minutes during and 5 minutes after insertion procedures Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: the swaddling + expressed breast milk method is clinically better pain relief non‐pharmacological option and is the recommended method to nurses for the orogastric tube insertion procedures, preterm infants Study funding source: Akdeniz University in Turkey Scientific Research Projects Coordination Unit and The Mustela Foundation Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Low risk	A convincing method for allocation concealment is reported.
3a) Performance bias (blinding of personnel)	High risk	Personnel were stated to be unblinded to group allocation.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were stated to be blinded.
4) Attrition bias (incomplete outcome data)	Low risk	Showed attrition reasons that appear balanced across groups.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Axelin 2006.

Study characteristics
Methods	Study design: randomised cross‐over trial Total study duration: 10 months (2003 to 2004)
Participants	Total number: 20 Setting: Turku University Hospital, NICU Diagnostic criteria: less than and equal to 37 weeks gestation, no major congenital anomalies, a need for regular endotracheal/pharyngeal suctioning, no analgesics for 4 hours before the procedure Age: Minimum: 24 weeks gestation, 6 days Maximum: 33 weeks gestation, 37 days Mean: not reported (median: 18 days) SD: not reported Sex: Males: 8 Females: 12 Country: Finland
Interventions	Total number of intervention groups: 1 Control group description: 1 Total number of participants randomised to each of the groups: 20 Specific intervention: facilitated tucking Intervention details: facilitated tucking by parents Integrity of intervention: satisfactory
Outcomes	Pain outcomes, time points when measured, and scale limits: NIPS Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that adverse events occurred. One infant got septicaemia after the experimental care and did not express pain during.
Notes	Funding sources: the South‐Western Finnish Foundation of Neonatal Research Key conclusions of study authors: facilitated tucking by parents is an effective and safe pain management method during suctioning of preterm infants Study funding source: The South‐Western Finnish Foundation of Neonatal Research Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported but this did not involve an independent person.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	Personnel could not be blinded but babies were.
3b) Detection bias (blinding of outcome assessors)	High risk	Coders could not be blinded to condition.
4) Attrition bias (incomplete outcome data)	High risk	There is no documented evidence or insufficient evidence reported of how attrition was dealt with.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (2/7), unclear risk (1/7).

Axelin 2009.

Study characteristics
Methods	Study design: randomised cross‐over Study duration: 2 years
Participants	Total number: 20 Setting: NICU at university hospital Diagnostic criteria: preterm, but were excluded if they had unstable health conditions Age: Minimum: 23 weeks Maximum: 30 weeks Mean: 28 SD: 2.3 Sex: Males: 12 Females: 8 Country: Finland
Interventions	Total number of intervention groups: 1 Control group description: sterile: 2 mL of water on tongue Total number of participants randomised to each of the groups: 20 Specific intervention #1: facilitated tucking Intervention details: parent held infant in a side lying, flexed fetal type position offering support and skin contact (taught parent the procedure in advance) Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: PIPP Time points: measured for the first 30 seconds Higher scores = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: facilitated tucking works Study funding source: funded by Turku University Hospital Research Foundation Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported but this did not involve an independent person.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Unclear risk	None specified.
3b) Detection bias (blinding of outcome assessors)	Low risk	A convincing effort to reduce bias in outcomes measurement is reported.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (0/7), unclear risk (2/7).

Ayazi 2017.

Study characteristics
Methods	Study design: between‐group Total study duration: 1 month
Participants	Total number: 112 Setting: hospital Diagnostic criteria: preterm Gestational age: Intervention: 34.4 weeks (2.0) Control: 34.6 weeks (2.1) Sex: Intervention: 28 males, 28 females Control: 31 males, 25 females Overall: 59 males, 53 females Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: in the control group, all measures were similar to the use of ear protectors except use of ear protector Total number of participants randomised to each of the groups: Intervention: 56 Control: 56 Specific intervention: ear protector Intervention details: prior to the intravenous sampling, infants wore the ear protector for 4 hours, after which it was removed. It was again restored when the intravenous sampling took place. Integrity of intervention: poor
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: pain scores were measured at two stages of intravenous sampling (pump time and at the moment of needle withdrawal) Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: the results of the study showed that the ear protector used for premature infants is effective in reducing pain during venipuncture Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Authors do not sufficiently describe the method used to generate the randomisation.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	It would be impossible to blind personnel to the intervention due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	It would be impossible to blind outcome assessors to the intervention due to nature of intervention.
4) Attrition bias (incomplete outcome data)	Low risk	Showed attrition reasons that appear balanced across groups.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	High risk	It appears that the use of incubators was not kept consistent across participants, which would introduce an additional source of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Aydin 2019.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 100 Setting: hospital Diagnostic criteria: full‐term, healthy Age: Mean: 39.1 (heel warming), 39 (control) SD: 1.1 (heel warming), 1.0 (control)   Sex: Males: 25 (heel warming), 25 (control) Females: 25 (heel warming), 25 (control) Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: in the control group, the heel stick procedure was conducted using the standard method, and the neonates received no intervention during the procedure. No nonpharmacologic methods for reducing pain during routine heel stick procedures were in routine use in the institution where the study was conducted. Total number of participants randomised to each of the groups: Heel warming: 50 Control: 50 Specific intervention: heel warming Intervention details: a thermal bag was used to heat the heel area before the heel stick procedure. The thermal bag is made of a nonleaking material with a fabric sheath, filled with warm water, and tightly closed. Superficial heat between 40°C and 45°C increases the blood flow. They put water at 40°C in the thermal bag and applied it against the puncture point for 3 to 5 minutes before the heel stick procedure. The heel stick procedure was performed immediately after removing the thermal bag. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: during procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: heel warming is effective in reducing the calming time during heel stick procedures Study funding source: this research did not receive any specific grant from funding agencies in the public, commercial, or not‐for‐profit sectors Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	High risk	Due to sequence generation method that was described, researcher could be aware of group assignment and allocation concealment would therefore not be maintained.
3a) Performance bias (blinding of personnel)	High risk	Personnel were not blinded to group allocation.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessor was stated to be blinded to group allocation.
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data observed.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Azarmnejad 2015.

Study characteristics
Methods	Study design: between‐group Total study duration: not mentioned
Participants	Total number: 30 Setting: hospital  Diagnostic criteria: term healthy neonates, lack of hearing impairment Gestational age (weeks):  Mean: 33.3 SD: 1.1 Sex: Males: 63.3% Females: 63.3% Country: Iran
Interventions	Total number of intervention groups: 2 Control group description: no pharmacological or non‐pharmacological interventions were applied Total number of participants randomised to each of the groups: Recorded maternal voice: 15 Control: 15 Specific intervention #1: mother's recorded voice Intervention details #1: mother's voice was recorded and played for the neonate by 2 small loudspeakers on both sides of the head at a distance of 20 cm from the neonate’s ear while he/she was in supine position on the baby cot in a separate room. The mother's voice was set at 50 dB to 60 dB by using a sound level meter, under the supervision of an audiologist. Then it was played for the intervention group 10 minutes before to 10 minutes after the intervention, in the morning  Integrity of intervention #1: good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: 10 minutes pre‐procedure, during and 10 minutes post‐procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors: intervention by mother group had the lowest level of pain compared to the others with the level found to be moderate with intervention by health workers. No significant differences in pain levels were found between the control and intervention by mother groups. It is recommended that facilitated tucking by mother combined with sitting up position should be used in infant pain management during intramuscular immunisation.  Study funding source: not reported Author COI statement: there are no conflicts of interest in this study
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Authors do not sufficiently describe the method used to generate the randomisation.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel unable to be blinded due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	It is not clear whether the outcome assessor was blinded to the intervention.
4) Attrition bias (incomplete outcome data)	Low risk	The number of enrolled participants in each group and the number reported in outcome tables match.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain is "high", the overall judgement is high.

Badiee 2014.

Study characteristics
Methods	Study design: between study Total study duration: 8 months
Participants	Total number: 100 Setting: hospital Diagnostic criteria: premature newborn twins Gestational age: Intervention: M = 32.1 weeks, SD = 1.8 Control: M = 32.0 weeks, SD = 1.7   Sex: Intervention: 24 females, 26 males Control: 23 females, 27 males Overall: 47 females, 53 males Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: remained in separate incubators following current NICU standards Total number of participants randomised to each of the groups: Intervention: 50 Control: 50 Specific intervention: co‐bedding Intervention details: infants in the co‐bedding group were placed side by side in an incubator without any clothing except for diapers so that they could touch each other freely, with each side of the incubator pertaining to one twin Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: before and 30 seconds after heel lance Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: co‐bedding could reduce pain sensation in premature infants as measured by the PIPP, and reduces cry time after heel lance Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	“Randomization was performed using a computer generated random number algorithm” (page 263, second paragraph of second column).
2) Selection bias (allocation concealment)	Low risk	“Allocation of eligible newborns to intervention and control groups was performed using a sealed opaque envelope” (page 263, second paragraph of second column).
3a) Performance bias (blinding of personnel)	Low risk	“Researchers could not be blinded for the assigned groups” (page 263, paragraph 5 second column).
3b) Detection bias (blinding of outcome assessors)	High risk	“PIPP score was calculated based on the recordings by three researchers.” (page 263, last paragraph second column) ‐ not sure if these are different researchers than those present for heel stick.
4) Attrition bias (incomplete outcome data)	Low risk	“…105 of them were included in the study. The main reason that infants could not be included was parental refusal. A few were excluded because there was no need for blood sampling” (page 264, paragraph 3 first column).
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score on Cochrane standard: low risk (6/7), high risk (1/7), unclear risk (0/7).

Bahorski 2015.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 173 Setting: hospital Diagnostic criteria: healthy, full‐term Age: 18 to 35 months Sex: Males: 96 Females: 77 Country: United States
Interventions	Total number of intervention groups: 3 Control group description: no control group Total number of participants randomised to each of the groups: LMX4 Cream = 66, Buzzy = 55, Buzzy + LMX4 = 52 Specific intervention: LMX4 Cream, Buzzy, Buzzy + LMX4 Intervention details: LMX4 Cream‐L.M.X.41 was placed on potential sites of venipuncture by the nurse or study personnel. The cream was left in place for a minimum of 20 minutes, then wiped off immediately before the venipuncture attempt. The child and parent were then immediately taken to the treatment room where the venipuncture procedure was conducted. The procedure was carried out per hospital procedure. A distraction method appropriate to the child’s age and of their choosing was utilised throughout the procedure. Buzzy: Buzzy was placed on the child proximal to the site of venipuncture, with the vibration component engaged, immediately before the venipuncture procedure, and left on during the entire procedure. This group also received distraction. Buzzy + LMX4: Buzzy was placed on the child proximal to the site of venipuncture, with the vibration component engaged, immediately before the venipuncture procedure, and left on during the entire procedure. This group also received distraction. L.M.X.41 was placed on potential sites of venipuncture by the nurse or study personnel. The cream was left in place for a minimum of 20 minutes, then wiped off immediately before the venipuncture attempt. Integrity of intervention: LMX4 Cream = good, Buzzy = good, Buzzy + LMX4 = good
Outcomes	Pain outcomes, scale limits, and time points when measured CHEOPS Time points: pre‐ and post‐treatment Upper limit: 13 Lower limit: 4 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: the authors conclude that, regardless of age or sex, both mechanical vibration and topical anaesthetic are effective in children Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	Mentions use of envelopes for randomisation but unclear whether they were sealed, opaque, or sequentially ordered. There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Unable to blind personnel to Buzzy condition.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	Unclear who completed the outcome assessment. It was scored in the child's hospital room, not the venipuncture room, so condition could have been blinded but if it was scored by parents, they would know what treatment their child received.
4) Attrition bias (incomplete outcome data)	Low risk	Reasons for dropout explained. Reasons for missing outcome data unlikely to be related to true outcome.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, an overall score of 'high' is given when one or more domain is determined to be high.

Basiri‐Moghadam 2014.

Study characteristics
Methods	Study design: between‐group Total study duration: May 2012 to February 2013 (9 months)
Participants	Total number: 34 Setting: clinic under the Gonabad University of Medical Sciences Diagnostic criteria: healthy Age Minimum: 4 months   Maximum: 6 months Mean: 7.4 (including EMLA group) SD: 1.0   Sex Males: N/A Females: N/A Country: Iran
Interventions	Total number of intervention groups: 2 Control group description: normal vaccination process Total number of participants randomised to each of the groups: Intervention 1: 16 Control: 18 Intervention details: rattles: a rattle was shaken from 30 seconds before administration to 15 seconds afterward as a source of distraction Integrity of intervention: good
Outcomes	Pain outcomes, time points when measured, and scale limits: Infants’ Pain Questionnaire Time points: infants were filmed before, during, and after injection (15 seconds). The observer used the films to fill out the questionnaires. Range of possible scores: the questionnaire scores facial expression and body motions in the range of 0 to 3, and at the range of 0 to 4 for crying. All the scores were summed. Maximum and minimum obtained scores of pain behavioural response were 10 and 0 respectively. Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: rattles reduced pain more than control, but EMLA was more significant compared to control during needle – NOTE: no significant differences between groups 15 seconds after Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	To ensure random selection, the researcher first chose a number out of a random number block and, depending on the right digit of the number, different combinations of words A (EMLA group), B for (rattle group), and C (control group) were adopted (1= ABC, 2 = BAC, 3 = CAB, 4 CBA, 5 = BCA, 6 = ACB; and re‐ selection of number for other digits (paragraph 5, page 875).
2) Selection bias (allocation concealment)	Unclear risk	No allocation concealment reported.
3a) Performance bias (blinding of personnel)	Low risk	To determine the severity of pain experienced by the infants and to blind and control the intervention factor, the infants were filmed before, during and after injection (15 seconds). The observer used the films to fill out the questionnaires. Only one individual was in charge of observation, who did not know about the classification of the participants (paragraph 5, page 875).
3b) Detection bias (blinding of outcome assessors)	Unclear risk	There was no report of blinding of outcome assessment.
4) Attrition bias (incomplete outcome data)	Low risk	Initially, 60 infants were recruited, but 14 (23.3%) subsequently had to be excluded; 2 (3.33%) did not meet the inclusion criteria, 4 (6.66%) were vaccinated by another person, and the questionnaire for 4 (6.66%) were not completely filled. The final study size comprised 50 (83.33%) infants (paragraph 3, page 876).
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (0/7), unclear risk (2/7).

Bauchner 1996.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 431  Setting: paediatric emergency department in a hospital Diagnostic criteria: infants aged 0 to 36 months coming to the paediatric ER, undergoing a venipuncture, intravenous cannulation, or urethral catheterisation Age (page 863): Minimum: newborn Maximum: 3 years Mean: unknown SD: unknown Sex: Males: 246 (57%) Females: 185 (43%) Country: USA Socio‐demographics: 87% of the mothers were between the ages 20 and 24 years old; 22% were married; 43% had less than a high school degree Ethnicity: Black: 249 participants Hispanic: 92 participants White: 35 participants Other: 39 participants Note: Although there was a total of 431 participants, ethnicity was only reported for 415 participants
Interventions	Total number of intervention groups: 2 Control group description: 1 (parents not present) Total number of participants randomised to each of the groups: 431 total (present with intervention: 153; parent presence 147; control: 131) Specific intervention #1: intervention ‐ parent instruction Intervention details: parents asked to sit at the head of the bed and talk to, touch, and maintain eye contact with their child Integrity of intervention: good Specific intervention #2: parent presence (treatment arm included only in qualitative analysis) Intervention details: parent present in the room, with no other instructions given Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured: Cry frequency (analyses) Measured during procedure (unspecified) in Hz. This involved the computerised analysis of cry, whereby each 30‐second cry signal was filtered above 10 kHz and digitised. In general, computerised analysis of cry confirmed reports of parents that the cry of children in pain is higher pitched and more turbulent. Higher frequency of cry = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: intervention (parent present, engaging in touch, talking and eye contact) during painful ER procedure are not effective in reducing pain (as measured by cry) in infants aged 0 to 36 months Study funding source: Maternal and Child Health Bureau; An Institutional Grant for Excellence in Pain Research from Bristol‐Myers Squibb Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Unclear risk	None specified.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (3/7), high risk (1/7), unclear risk (3/7).

Baudesson 2017.

Study characteristics
Methods	Study design: between‐group Total study duration: 23 months
Participants	Total number: 33 Setting: hospital (NICU) Diagnostic criteria: stable preterm infants Gestational age: Mean: 32.4 (maternal milk odour), 33.6 (control), 33.2 (overall) SD: none reported Sex: Males: 15 Females: 18 Country: France
Interventions	Total number of intervention groups: 1 Control group description: all the neonates were given a dummy. The use of a dummy was identical in the two groups, thus allowing the non‐pharmacologic analgesic method of non‐nutritive sucking in the control group. A unit professional nurse performed the venipuncture with a 20 G needle while following a standardised procedure. Total number of participants randomised to each of the groups: Maternal milk odour: 16 Control: 17 Specific intervention: maternal milk odour Intervention details: all the neonates were given a dummy. The use of a dummy was identical in the 2 groups, thus allowing the non‐pharmacologic analgesic method of non‐nutritive sucking in the control group. An air‐flow odour diffuser was filled with 5 g of a water‐soluble polymer gel in the form of beads that absorbed and then slowly released an odour over time. The diffuser was placed under a hood with an air flow of 6 L/minute and was manually switched on 3 minutes before the venipuncture and switched off 9 minutes after the completion of the procedure. A picture of the device is shown in Photograph 1. The type and quantity of maternal milk to be used were determined before the study by an olfactory test. A unit professional nurse performed the venipuncture with a 20 G needle while following a standardised procedure. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: maternal milk odour has an analgesic effect on preterm neonates Study funding source: this work was supported by institutional grants from the French 2013 Appel d’Offre Recherche Clinique Assistance Publique, Hôpitaux de Marseille Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	Low risk	Personnel were blinded to study condition.
3b) Detection bias (blinding of outcome assessors)	Low risk	Coding done from videos so blinding was possible.
4) Attrition bias (incomplete outcome data)	Low risk	No attrition reported.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, an overall score of 'unclear' is given when one or more domain is determined to be unclear.

Behnam Vashani 2015.

Study characteristics
Methods	Study design: between‐group Total study duration: 2 months
Participants	Total number: 80 Setting: hospital Diagnostic criteria: preterm neonates Fetal age: Mean: 30.4 weeks (multi‐sensory bundle), 30.6 weeks (control) SD: 1.7 (multi‐sensory bundle), 1.8 (control) Chronological age: Mean: 35.1 weeks (multi‐sensory bundle), 34.9 weeks (control) SD: 1.9 (multi‐sensory bundle), 1.8 (control) Sex: Males: 37 Females: 43 Country: Iran
Interventions	Total number of intervention groups: 1 Description of control: the eye examinations were performed within an hour after breastfeeding by the mother while the baby was calm and sober. The control group received no interventions and were provided with similar care to the intervention group. Total number of participants randomised to each of the groups: Control: 40 Multi‐sensory bundle: 40 Intervention descriptions: Specific intervention: multi‐sensory bundle Description of intervention: as the facial tactile stimuli, the upper and lower limbs of the neonates were gently touched by the mother. As for the visual stimulation, the mother would look at the neonate’s face trying to make close eye contact and attract the attention of the baby. These interventions were performed 15 minutes prior to the beginning of the examinations. For auditory stimulation, the mother would speak to the neonate gently and continuously for as long as 15 minutes prior to the examinations. As for the olfactory stimulation, we used vanilla solution and stained a piece of sterile gauze. The gauze was held at a distance of approximately 1 mm to 2 mm from the baby’s nose without any contact. Similarly, this intervention was performed 15 minutes prior to the beginning of the eye examinations. For the gustatory stimulation, we used 1 mL of 33% glucose solution pulled by a syringe by the researcher and placed in the infant's mouth without any needles. As the infant gently sucked the syringe, the solution was injected into the mouth for 30 seconds. This intervention was performed 2 minutes prior to the eye examinations. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP (only used facial expression subscales) Time points: first 30 seconds post‐procedure and 1 minute post procedure Upper limit: 9 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: according to the results of this study, multi‐sensory stimulation was able to reduce the manifestations of facial expression in the studied neonates Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Authors do not sufficiently describe the method used to generate the randomisation.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	The person performing the eye exam was unable to be blinded to study intervention.
3b) Detection bias (blinding of outcome assessors)	Low risk	Given that interventions were applied prior to the procedure, blinding was possible.
4) Attrition bias (incomplete outcome data)	Unclear risk	Insufficient reporting of attrition/exclusions to permit judgement of risk of bias (e.g. number randomised not stated, no reasons for missing data provided). It was unclear how missing participants or data were accounted for.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under high risk then the overall assessment is high risk.

Beiranvand 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 120 Setting: hospital Diagnostic criteria: healthy full‐term Age (months): Mean: 3.7 (hugging and caressing), 3.7 (control), overall mean = 3.7 SD: 1.7 (hugging and caressing), 1.7 (control), overall SD = 1.4   Sex: Males: 30 (hugging and caressing), 30 (control), total = 60 Females: 30 (hugging and caressing), 30 (control), total = 60 Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: infants were placed on bed 2 minutes before and after venipuncture (according to the usual venipuncture method). 15 minutes before venipuncture, infants were fed with breast milk or formula. Total number of participants randomised to each of the groups: Intervention: 60 Control: 60 Specific intervention: hugging and caressing Intervention details: infants were placed on the mothers’ lap for 2 minutes before and after venipuncture and caressed by the mother during this period. 15 minutes before venipuncture, infants were fed with breast milk or formula. Integrity of intervention: poor
Outcomes	Pain outcomes, scale limits, and time points when measured MBPS Time points: pain measured before, during catheter insertion and after venipuncture Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: overall, caressing in infants had positive effect on pain relief after venipuncture Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	There is not an adequate description of the method utilised for generation of a random allocation sequence.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel unable to be blinded due to nature of intervention (holding by mother versus on table).
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors unable to be blinded due to nature of intervention.
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data observed.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Bellieni 2001.

Study characteristics
Methods	Study design: cross‐over Total study duration: 3 months (December 2000 to February 2001)
Participants	Total number: 17 Setting: Siena Hospital NICU Diagnostic criteria: preterm infants Age: all infants were born less than 35 weeks gestation, and the procedure took place within 10 days post natal life Minimum: 28 weeks Maximum: 35 weeks Mean: unknown SD: unknown Sex: Males: 7 Females: 10 Country: Italy
Interventions	Total number of intervention groups: 1 Control group description: no intervention whatsoever (just heel lance) Total number of participants randomised to each of the groups: 17 Specific intervention: water was administered orally 30 seconds before and during withdrawal Intervention details: water was instilled on the tip of the tongue with a syringe, introduced in mouth and moved to stimulate sucking. The amount that was administered was that necessary to maintain sucking (0.2 mL to 0.3 mL) until after the heel prick. Integrity of intervention: satisfactory
Outcomes	Pain outcome, scale limits, and time points when measured: PIPP Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: non‐nutritive sucking is effective in reducing pain for preterm infants undergoing heel lance when compared to receiving no treatment at all Miscellaneous comments by review authors: poor description of methods and time in which pain was measured; poor description of sample characteristics Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported but this did not involve an independent person.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Unclear risk	A convincing effort to reduce bias.
3b) Detection bias (blinding of outcome assessors)	High risk	Coders not blinded to receiving no treatment or water.
4) Attrition bias (incomplete outcome data)	High risk	There is no documented evidence or insufficient evidence reported of how attrition was dealt with.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (2/7), high risk (3/7), unclear risk (2/7).

Bellieni 2002.

Study characteristics
Methods	Study design: between‐group Total study duration: 5 months (February to June 2001)
Participants	Total number: 60  Setting: hospital Diagnostic criteria: healthy newborns Age: Minimum: 38 weeks Maximum: 41 weeks Mean: not reported SD: 2.5 Sex: Males: 31 Females: 29 Country: Italy Ethnicity: Caucasian
Interventions	Total number of intervention groups: 2 Control group description: no analgesic procedure Total number of participants randomised to each of the groups: 20 in each group (total of 60) Specific intervention #1: water during heel prick (sucking) Intervention details: the tip of 1 mL syringe without needle was placed in the baby's mouth and 1 mL of distilled water was given with gentle a movement of the syringe to stimulate sucking for 30 seconds before, during, and after heel stick Integrity of intervention: good Specific intervention #2: sensorial saturation without glucose (treatment arm included only in qualitative analysis) Intervention details: involved laying the infant on its side with legs and arms flexed but free to move; looking at the infant in the face, close up, to attract its attention and simultaneously massaging the infant's face and back; speaking to the infant softly but firmly; letting the infant smell the fragrance of the baby perfume on the physiotherapist's hands Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: Douleur Aiguë Nouveau‐né (DAN) scale Time points: 30 seconds pre heel stick, 30 seconds post heel stick Upper limit: 10 Lower limit: 0 Higher score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: sensorial stimulation without glucose does not produce an analgesic effect in healthy newborns during the 30‐second post‐heel prick but rather, it increased irritation and awareness Sucking (water oral) does significantly reduce pain scores during 30 seconds post‐heel lance in healthy newborn infants Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	A convincing effort to reduce bias in outcomes measurement is reported.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (0/7), unclear risk (2/7).

Bembich 2018.

Study characteristics
Methods	Study design: between‐group Total study duration: 7 months
Participants	Total number: 40 Setting: hospital Diagnostic criteria: healthy full‐term neonates (3rd day of life) Gestational age: Mean: 39.8 (glucose + maternal holding), 39.6 (control) SD: none reported Sex: Males: 18 Females: 22 Country: Italy
Interventions	Total number of intervention groups: 1 Control group description: the infants received 2 mL of oral glucose solution 2 minutes before the heel stick, which was performed on a changing table. Infants who were allocated to group one were placed on a changing table, and fibres were positioned on the scalp. A waiting period was allowed for the infant to get used to the equipment. Two minutes before starting the heel stick procedure, a 2 mL bolus of 20% oral glucose solution was administered directly into the infant’s mouth with a syringe. Total number of participants randomised to each of the groups: Glucose + maternal holding: 20 Control: 20 Specific intervention: glucose + maternal holding Intervention details: infants were held in their mothers’ arms throughout the procedure and received 2 mL of oral glucose solution 2 minutes before the heel stick. Infants in group 3 were tested while in their mothers’ arms (mother–infant relationship). Optical fibres were placed on the scalp, and a waiting period was allowed for the newborn to adapt. Two minutes before the heel stick procedure, a 2 mL bolus of 20% oral glucose solution was given directly into the infant’s mouth with a syringe. The mothers were asked not to talk to their infants. Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: first 30 seconds post‐procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: pain expression was lower with the maternal‐infant relationships Study funding source: supported by a grant from the Institute for Maternal and Child Health–IRCCS “Burlo Garofolo”–Trieste (N. 50/11) Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes potential biases of investigators.
3a) Performance bias (blinding of personnel)	High risk	Nurse performing heel stick not blinded.
3b) Detection bias (blinding of outcome assessors)	High risk	Although NIPS outcome assessors were unaware of the cortical activity of infants, they could not have been blinded to the intervention (could see if mother was holding the baby or not, or if the baby was being breast fed during procedure). Also NIPS scoring was not done from video, so outcome assessor was in the room.
4) Attrition bias (incomplete outcome data)	Low risk	Authors provide flow chart, and it seems there were no missing outcome data.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall assessment is high risk.

Benjamin 2016.

Study characteristics
Methods	Study design: between group Total study duration (months): 8.2
Participants	Total number: 100 Setting: publicly funded, free immunisation clinic and a private group paediatric practice Diagnostic criteria: healthy children Postnatal age: 2 months to ≤ 1 year Mean: not mentioned  SD: not mentioned  Sex: Males: 52 Females: 48 Country: USA
Interventions	Total number of intervention groups: 2  Control group description: the control group received the scheduled immunisation injection without vibration. This group was allowed to be held, soothed, comforted, and distracted by both the nurse and parent, as would have been normally done. Total number of participants randomised to each of the groups: Vibration: 50 Control: 50 Specific intervention #1: vibration   Intervention details #1: this group received vibration via the Buzzy vibrating device 10 seconds prior to receiving the injection. The device was placed anatomically superior to the injection site, along the same skin dermatome. The nurse injected the immunisation within approximately 0.5 cm from the device. The vibration device remained vibrating next to the skin while the injection was administered until after the needle was removed. This group was also allowed to be held, soothed, comforted, and distracted by both the nurse and parent, as would have been done normally. Integrity of intervention #1: good
Outcomes	Pain outcomes, scale limits, and time points when measured FLACC Time points: 30 seconds prior to, during, and 30 seconds after the procedure Upper limit: 10  Lower limit: 0 High score = more pain Continuous Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors: vibration therapy had no effect on pain scores in the younger age groups studied (2 months ≤ 1 year, > 1 year ≤ 4 years). These results indicate that the addition of vibration therapy (without cold analgesia) to standard soothing techniques is no more effective in reducing immunisation pain than standard soothing techniques alone and, thus, is not indicated for use with immunisation pain.  Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel unable to be blinded due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors unable to be blinded due to nature of intervention.
4) Attrition bias (incomplete outcome data)	Low risk	The number of enrolled participants in each group and the number reported in outcome tables match.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	High risk	The study was conducted at 2 separate clinics that employed different immunisation procedures, which could have introduced systematic differences across participants.
Total	High risk	As per Sterne 2019, if one or more domain is "high", the overall judgement is high.

Blass 1999.

Study characteristics
Methods	Study design: randomised between groups Total study duration: 12 months (June 1997 to June 1998)
Participants	Total number: 20 Setting: hospital (Boston Medical) Diagnostic criteria: healthy newborns Age: Minimum: 34 hours old Maximum: 55 hours old Mean: not reported SD: not reported Sex: Males: 17 Females: 23 Country: USA Ethnicity: 55% African American; 22.5% White; 12.5% Hispanic; 2.5% Asian; 2.5% American Indian; 5% unclassified
Interventions	Total number of intervention groups: 1 ‐ pacifier dipped in water Control group description: water only (delivered by syringe) Total number of participants randomised to each of the groups: 20 total (10 in each group) Specific intervention #1: infant was given either water only or a pacifier prior to heel lance Intervention details: 4 minutes pre heel lance, the infant was given a heat pad to warm feet and was wrapped in a blanket and turned supine to his or her bassinet. 60 seconds before the heel stick, the infant was given either water only (delivered by syringe) or pacifier (coated in water). Pacifier was re‐dipped in water every 30 seconds. The heat pad was then removed, and the heel lance occurred lasting 1 to 3 minutes. A band‐aid was then placed on the wound. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: Crying Time points: heel lance (2 minutes post), recovery (5 minutes post) Upper limit: 100% Lower limit: 0% High score = more pain Continuous Grimacing (furrowing of brow and eye squinting) Time points: heel lance (2 minutes post), recovery (2 minutes post) Upper limit: 100% Lower limit: 0% Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: sucking a water‐dipped pacifier caused modest and variable reductions in grimacing during the 2‐minute post‐heel lance in healthy neonates. Further analyses revealed that this intervention is only effective if the sucking rate is greater than 30 times/minute prior to the heel lance. Miscellaneous comments by review authors: results are not significant, and no data are given at the recovery period (2 minutes to 5 minutes post heel lance) Study funding source: this research has been supported by Grant RO‐1 MH51705‐04A1 and Research Scientist Award KO‐5, MH5 00524, from the National Institutes of Mental Health to EMB Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported but this did not involve an independent person.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Low risk	A convincing effort to reduce bias in outcomes measurement is reported.
3b) Detection bias (blinding of outcome assessors)	High risk	Not blinded to pacifier versus syringe condition as seeing the face for grimacing would have involved seeing the pacifier in the infant's mouth.
4) Attrition bias (incomplete outcome data)	High risk	There is no documented evidence or insufficient evidence reported of how attrition was dealt with.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (3/7), high risk (2/7), unclear risk (2/7).

Bo 2000.

Study characteristics
Methods	Study design: cross‐over Total study duration: not reported
Participants	Total number: 27 Setting: special care baby unit, NICU Diagnostic criteria: healthy neonates and premature neonates Age: Minimum: 30 weeks Maximum: 41 weeks Mean: 37 weeks SD: 3.4 weeks Sex: Males: 17 Females: 10 Country: Hong Kong/China Ethnicity: Chinese
Interventions	Total number of intervention groups: 1 Control group description: infants were placed in the supine position. Total number of participants randomised to each of the groups: 27 Specific intervention: non‐nutritive sucking Intervention details: infants were given a small, standard, short and hollow soft latex nipple. Gentle pressure was applied to keep it in the infant's mouth. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: NIPS Time points: 1 minute pre heel stick, 13 minutes post heel stick Upper limit: 7 Lower limit: 0 Higher score = higher pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: non‐nutritive sucking is an effective intervention for pain regulation after heel stick in neonates Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported, but this did not involve an independent person.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	A convincing effort to reduce bias in outcomes measurement is reported.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	High risk	There is no documented evidence or insufficient evidence reported of how attrition was dealt with.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (1/7), unclear risk (2/7)

Bustos 2008.

Study characteristics
Methods	Study design: randomised between groups Total study duration: not reported
Participants	Total number: 50 Setting: Sydney Children's Hospital Outpatient Immunization Clinic Diagnostic criteria: adequate grasp of English, spoke English to infant Age: Minimum: 5 months Maximum: 7 months Mean: not reported SD: not reported Sex: Males: 24 (in the final sample) Females: 26 (in the final sample) Country: Australia
Interventions	Total number of intervention groups: 1 Control group description: standard care Total number of participants randomised to each of the groups: Intervention: 25 Control group: 25 Specific intervention: coping ‐ promoting verbalisations Intervention details: intervention group received a one‐page information sheet with information about 3 types of parental verbalisations associated with lower pain outcomes for infants. Parents were encouraged to review the sheet before their appointment. All infants were placed in supine position on the treatment table. The injection site was cleaned with an alcohol swab and the injection was then performed. During the injection, parents held infant's leg. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: NFCS Time points: injection phase: 15 seconds post needle, recovery phase: 15 seconds following injection phase (scored in 5 second blocks) Upper limit: 30 Lower limit: 0 High score = higher expression of pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: the promotion of parent‐coping verbalisations during an infant's 6‐month immunisation is able to reduce the duration of infant cry following an injection There was no significant effect of intervention on NFCS outcome measure Study funding source: The Pain in Child Health Strategic Training Initiative in Health Research, Canadian Institutes of Health Research, and the Mayday Foundation provided fellowship support Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	High risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	Low risk	All coders were blinded to group assignment.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (6/7), high risk (1/7), unclear risk (0/7).

Caglar 2017.

Study characteristics
Methods	Study design: between‐group Total study duration: 12 months
Participants	Total number: 100 Setting: hospital Diagnostic criteria: healthy full‐term neonates Gestational age: Mean: 39.1 (intervention), 38.9 (control) SD: 0.9 (intervention), 0.9 (control) Sex: Males: 56 Females: 44 Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: neonates received swaddling by a soft cotton hospital blanket. First, the neonates were placed in a supine position on a blanket while their arms were flexed, crossed over, and close to the chest. Then, a nurse researcher administered the vaccine within the first 15 minutes of delivery. The skin was cleansed with an alcohol swab before injection and allowed to dry for 30 seconds. The vaccine was then given by IM injection in the vastus lateralis muscle with a 1‐inch, 23‐gauge needle inserted at a 90º angle while grasping the tissue. After aspirating to confirm the needle had not been placed in a vein or artery, the vaccine was injected. Total number of participants randomised to each of the groups: Control: 50 ShotBlocker + swaddling: 50 Specific intervention: ShotBlocker + swaddling Intervention details: on top of swaddling, while receiving the injection, this group had the ShotBlocker placed at the injection site. The ShotBlocker was held firmly against the skin for 20 seconds. The injection was administered as usual. ShotBlocker was removed after needle was withdrawn. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: during procedure and 3 minutes after Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: the use of ShotBlocker during the hepatitis B vaccine in term neonates is effective in reducing the acute pain Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel performing the immunisation not blinded.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors were not blinded to intervention.
4) Attrition bias (incomplete outcome data)	Low risk	Authors provide flow chart and explain reasons for missing outcome data, which appear balanced across groups.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall assessment is high risk.

Campbell‐Yeo 2012.

Study characteristics
Methods	Study design: between‐group Total study duration: 28 months
Participants	Total number: 67 sets of twins (124 infants) Setting: NICU Diagnostic criteria: eligibility was based on being preterm (28 to 36 weeks gestation), heavier than 1000 g, without major anomalies, requiring at least one medically indicated heel lance, and medically stable Age: Mean: 31.9 (gestational age) SD: 2.6 Sex: Males: 76 Females: 58 Country: Canada
Interventions	Total number of intervention groups: 1 Control group description: standard care: remained in separate incubator. Both the intervention and standard care groups received sucrose 2 minutes before heel lance as part of hospital policy. Total number of participants randomised to each of the groups: Intervention: 67 Control group: 57 Specific intervention: co‐bedding Intervention details: placed in diaper clad together in giraffe incubator close to each other, permitting touch Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: PIPP Time points: 30 seconds epochs from the time of the heel lance until the application of the bandage Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: co‐bedding enhanced the physiologic recovery of preterm twins undergoing heel lance, but did not lead to lower pain scores Study funding source: partnered funding for this study was obtained from the Groupe de Recherche Interuniversitaire en Interventions en Sciences Infirmieres du Quebec, the Canadian Nurses Foundation Nursing Care Partnership Program, CIHR, and the IWK Health Centre Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	High risk	Participants (i.e. parents) and personnel could not be blinded due to seeing twins in either co‐bedding arrangement or standard care, “Limitations in our study include the inability to blind care providers to the intervention” (p. 505).
3b) Detection bias (blinding of outcome assessors)	Low risk	Coders were not aware of the purpose of the markers or study objective.
4) Attrition bias (incomplete outcome data)	Low risk	Information included on data in a flow chart.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (6/7), high risk (1/7), unclear risk (0/7).

Campos 1994.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 60 Setting: mid‐western neo‐natal nursery Diagnostic criteria: full‐term birth; Apgar scores > 7; absence of respiratory or other health problems Age: Minimum: not reported Maximum: not reported Mean: 51.5 hours SD: 12 hours Sex: Males: 30 Females: 30 Country: USA
Interventions	Total number of intervention groups: 2 Control group description: infants received no comforting care after heel stick; were placed prone in bassinet after heel stick and covered with a blanket Total number of participants randomised to each of the groups: 20 participants in each of the 3 groups, total number of participants is 60 Specific intervention #1: rocking Intervention details: infants were held at the experimenter's shoulder, rocked in a rocking chair at a rate of 30 cycles per minute (auditory signal fed to experimenter) post‐procedure Integrity of intervention: good Specific intervention #2: pacifier Intervention details: infants were held in the experimenter's lap while the pacifier was administered for 8 minutes after heel stick; gentle pressure was applied to keep that pacifier in the infant's mouth Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: Latencies to cry cessation Time points: 0 to 2 minutes post heel stick, 3 to 4 minutes post heel stick Upper limit: 180 seconds Lower limit: 0 seconds Higher score = longer latency to cry cessation Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: newborns benefit from both rocking and pacifiers when undergoing heel stick procedures Study funding source: this research was partially supported by grants from the National Association of Neonatal Nurses and the University of Illinois Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported, but this did not involve an independent person.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	A convincing effort to reduce bias in outcomes measurement is reported.
3b) Detection bias (blinding of outcome assessors)	Low risk	Coders were blinded to group assignment.
4) Attrition bias (incomplete outcome data)	High risk	There is no documented evidence or insufficient evidence reported of how attrition was dealt with.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (0/7), unclear risk (2/7).

Carbajal 1999.

Study characteristics
Methods	Study design: between‐group Total study duration: April 1997 to June 1997
Participants	Total number: 50  Setting: maternity ward of a hospital in France Diagnostic criteria: newborns who had a 5‐minute Apgar score greater than or equal to 7, were medically stable, had not received naloxone during the previous 24 hours, and were not fed in the previous 30 minutes Age: Minimum: 37 weeks Maximum: 42 weeks Mean: 40 weeks SD: not reported Sex: Males: 89 Females: 61 Country: France
Interventions	Total number of intervention groups: 5 (but only one does not use sucrose) Control group description: placebo, 2 mL of sterile water Total number of participants randomised to each of the groups: Intervention: 25 Control: 25 Specific intervention #1: pacifier Intervention details: pacifier given to baby Integrity of intervention: unknown
Outcomes	Pain outcomes, scale limits, and time points when measured: DAN (Douleur Aiguë Nouveau‐né) Time points: not mentioned Upper limit: 10 Lower limit: 0 Higher score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: pacifiers were effective in reducing infant pain Study funding source: no external funding Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	High risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None provided.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (1/7), unclear risk (2/7).

Carbajal 2003.

Study characteristics
Methods	Study design: between‐group Total study duration: 5 months
Participants	Total number: 180 (90 included in quantitative analysis) Setting: maternity ward of hospital Diagnostic criteria: equal to or greater than 37 weeks gestation; had Apgar scores of 7 or higher at 5 minutes; were greater than or equal to 24 hours old; undergoing venipuncture; breast‐fed; had not been fed for the previous 30 minutes; no medical instability; no medication in the previous 24 hours Age: Minimum: 39.6 weeks Maximum: 40 weeks Mean: 39.8 SD: 1.2 Sex: Males: 93 Females: 87 Country: France
Interventions	Total number of intervention groups: 3 (we will only focus on the first one for this review; other interventions included breast milk and sucrose) Control group description: infants given sterile water as a placebo Total number of participants randomised to each of the groups: 45 in each group (180 total) Specific intervention: held in mother's arms without breastfeeding Intervention details: held in mother's arms, without breastfeeding Integrity of intervention: satisfactory
Outcomes	Pain outcome, scale limits and time points when measured: Douleur (DAN) scale Time points: unknown Upper limit: 10 Lower limit: 0 High score = more pain Continuous PIPP Time points: unknown Upper limit: 21 Lower limit: 0 Higher score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: there were no significant differences in pain reactivity between newborns being held in mother's arms or newborns who were given sterile water Study funding source: Fondation pour la Santé CNP, France Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	High risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	High risk	Coders were not blinded.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted, so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (2/7), unclear risk (1/7).

Catelin 2005.

Study characteristics
Methods	Study design: cross‐over Total study duration: not reported
Participants	Total number: 45 Setting: NIDCAP‐reliable NICU Diagnostic criteria: participants were excluded if they received treatment with a muscle relaxant, sedative, antiepileptic, or analgesic drug (except sucrose) during the last 24 hours; had a congenital defect, a neurological abnormality including convulsion, intraventricular haemorrhage grade higher than II according to the Papile scale, and periventricular leukomalacia Age: Minimum: less than or equal to 32 weeks Maximum: greater than or equal to 37 weeks Mean: 34.5 weeks SD: 1.0 weeks Sex: Males: 27 Females: 18 Country: France
Interventions	Total number of intervention groups: 1 Control group description: no EBI, no specific protection from light or noise, in the supine position and without swaddling or any postural support Total number of participants randomised to each of the groups: 45 Specific intervention: environmental and behavioural interventions (EBI) Intervention details: EBI included attenuated noise and light with closed doors and covered incubator, lateral posture with head, back, and feet contacting supportive bedding, and opportunity for grasping or sucking Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured: NIPS Time points: 2 minutes before weighing, during weighing and 5 and 30 minutes after weighing Upper limit: 7 Lower limit: 0 High score = more pain Continuous EDIN 2 minutes before and 5 and 30 minutes after weighing Upper limit: 15 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: EBI procedures were associated with lower heart rates just after weighing procedures Study funding source: supported by grants from the French Ministry of Health (PHRC 2002), the Fondation de France, SESEP and the Fondation CNP Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	A convincing effort to reduce bias in outcomes measurement is reported.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (0/7), unclear risk (2/7).

Cetinkaya 2020.

Study characteristics
Methods	Study design: between group Total study duration: 7 months
Participants	Total number: 60 Setting: hospital (NICU) Diagnostic criteria: preterm and full‐term Age: 46.7% preterm, 53.3% full‐term Sex: 29 female, 31 male Country: Turkey
Interventions	Total number of intervention groups: 1   Control group description: newborns in the control group listened to white noise after videotaping. A CD of white noise was given to mothers of newborns in the control groups. Total number of participants randomised to each of the groups: Intervention: 30 Control: 30 Specific intervention: white noise Intervention details: white noise was listened to just before the invasive procedure (5 to 10 seconds before) and it was continued until the procedure was finished even if newborn stopped crying. White noise was turned off when the procedure was finished (in the first 30 seconds). This CD was created and used by Celik and Cetinkaya (2015) for their Postgraduate Thesis, and included the noises of a hairdryer, vacuum cleaner, and heartbeats sound like music, and made newborn babies feel safe and comfortable. Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured   NIPS   Time points: during procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: the use of white noise CD on newborns during invasive procedures has been determined to be applicable Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	The authors state the sequence was generated through "simple random sampling method", which is not enough information to allocate a risk of bias rating.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Because infants in the control group listened to white noise during the procedure, all personnel would not be blinded to intervention allocation.
3b) Detection bias (blinding of outcome assessors)	High risk	Unable to blind outcome assessors because crying is evaluated.
4) Attrition bias (incomplete outcome data)	Low risk	60 infants were recruited and data on 60 infants are reported.
5) Selective reporting (reporting bias)	Unclear risk	Trial register not reported so unable to assess whether planned outcomes align with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall assessment is high risk.

Chang 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: 24 months
Participants	Total number: 101 Setting: hospital Diagnostic criteria: full‐term, healthy Age: 24 to 48 hours Sex: not reported Country: United States
Interventions	Total number of intervention groups: 1 Control group description: newborns were placed in a crib in supine position and covered with blankets Total number of participants randomised to each of the groups: Non‐nutritive sucking: 51 Control: 50 Specific intervention: non‐nutritive sucking Intervention details: infants received pacifiers before and during the procedure Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured N‐PASS Time points: duration of the procedure was calculated from the point of lancet puncture until blood collection was completed. N‐PASS scores were collected at these same procedure points. Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: non‐nutritive sucking has analgesic effects that can decrease the amount of pain that newborns experience Study funding source: Lucile Packard Foundation Children’s Health Fund and the Stanford NIH CTSA Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel were stated to be unblinded to group allocation.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors were stated to be unblinded to group allocation.
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data observed.
5) Selective reporting (reporting bias)	Unclear risk	While a trial register was reported, the study was retroactively registered so unable to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Chen 2017.

Study characteristics
Methods	Study design: between‐group Total study duration: 3 months
Participants	Total number: 26 Setting: hospital (NICU) Diagnostic criteria: 80% preterm Gestational age: Mean: 34.1 (magnet), 34.4 (control) SD: not reported Postnatal age: Mean: 9.6 days (magnet), 6.5 days (control) SD: not reported Sex: Males: 21 Females: 19 Country: Australia
Interventions	Total number of intervention groups: 1 Control group description: placebo stickers that were applied bilaterally to the 5 auricular acupuncture points as per the Battlefield Acupuncture (BFA) protocol at least 2 hours prior to a heel prick by the unblinded investigator. Each heel prick followed standard routine in accordance with the Royal Hospital for Women NICU guidelines. All infants received routine care with sucrose administered prior to heel pricks, and further analgesia provided according to clinical judgement (including nurse‐directed sucrose, kangaroo care, breastfeeding, and swaddling). No opioids were used as analgesia for heel pricks. Total number of participants randomised to each of the groups: Magnet: 11 Control: 15 Specific intervention: magnet Intervention details: MA stickers placed at same site as placebo stickers. Magnets had a strength of 100 G and measured 1.7 mm in diameter on a circular sticker of surgical tape with diameter 10 mm. Each heel prick followed standard routine in accordance with the Royal Hospital for Women NICU guidelines. All infants received routine care with sucrose administered prior to heel pricks, and further analgesia provided according to clinical judgement (including nurse‐directed sucrose, kangaroo care, breastfeeding, and swaddling). No opioids were used as analgesia for heel pricks. Integrity of intervention: poor
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: this pilot study shows that auricular MA is feasible in neonates and may reduce PIPP scores during heel pricks Study funding source: no specific funding Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	Due to the use of placebo stickers with magnets removed at the exact same sites as the magnetic stickers, personnel were blinded.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were blinded to study condition.
4) Attrition bias (incomplete outcome data)	Low risk	Attrition rates explained properly. Reasons for attrition were similar across groups.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Low risk	As per Sterne 2019, if all domains fall under low risk, the overall assessment is low risk.

Chen 2019.

Study characteristics
Methods	Study design: between‐group Total study duration: 5 months
Participants	Total number: 72 Setting: hospital (NICU) Diagnostic criteria: stable preterm neonates Gestational age: Mean: 34 (maternal voice stimulation), 35 (control) SD: not reported Sex: Males: 40 Females: 32 Country: China
Interventions	Total number of intervention groups: 1 Control group description: the heel blood collection of the control group was carried out according to the operation routine Total number of participants randomised to each of the groups: Maternal voice stimulation: 35 Control: 37 Specific intervention: maternal voice stimulation Intervention details: on top of the routine operation, mother sound will be recorded 1 minute before blood collection. The voice recorder is placed in the newborn warm box to keep the sound volume within 50 dB, until the end of the operation, the playback time is about 10 minutes. The mother’s voice is in a song, a story or a speech recorded by the newborn mother in a quiet space. The recording is about 5 minutes long. Mother can select the song, story or speech content to record. Researcher provides recording reference materials. All operations are recorded under camera. Nurse collect blood sample and records the heart rate, respiration and oxygen saturation levels in the blood of newborns from 30 seconds to 5 minutes before and after puncture. Two researchers playback recorded video after all procedures have been done, and do pain score at each time point (before and after puncture). Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: 30 seconds post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: maternal voice stimulation helps to reduce procedural pain in neonates
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	Unclear risk	It is not clear whether the personnel were blinded to the intervention.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	It is not clear whether the outcome assessor was blinded to the intervention.
4) Attrition bias (incomplete outcome data)	Unclear risk	Attrition/exclusions not described so unable to determine whether there was incomplete outcome data.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	High risk	They included both preterm and term babies that resulted in an age difference within the two groups.
Total	High risk	As per Sterne 2019, given that there is a rating of high risk for one of the domains, the overall rating is high.

Chik 2012.

Study characteristics
Methods	Study design: cross‐over Total study duration: 7 months
Participants	Total number: 65 Setting: NICU Diagnostic criteria: preterm infants in NICU Age: Minimum: unknown Maximum: unknown Mean: unknown SD: unknown Sex: Males: unknown Females: unknown Country: Hong Kong
Interventions	Total number of intervention groups: 2 Control group description: usual care Total number of participants randomised to each of the groups: 65 Specific intervention: limb massage Intervention details: 2‐minute limb massage prior to procedure and vice versa Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: right after the intervention and during the first 30 seconds of the procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: upper limb massage may be effective technique for relieving infants' venipuncture pain Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Not enough information provided in abstract.
2) Selection bias (allocation concealment)	Unclear risk	Not enough information provided in abstract.
3a) Performance bias (blinding of personnel)	Unclear risk	None specified.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Unclear risk	Not enough information provided in abstract.
5) Selective reporting (reporting bias)	Unclear risk	Not enough information provided in abstract.
6) Other bias	Unclear risk	Not enough information provided.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (0/7), high risk (0/7), unclear risk (7/7).

Chik 2017.

Study characteristics
Methods	Study design: cross‐over Total study duration: not reported
Participants	Total number: 80 Setting: NICU Diagnostic criteria: sample included both premature and full‐term infants (gestational age range at birth 30 to 40 weeks). Venipuncture to test for infection suggesting overall healthy populations. Gestational age: 37.3 (upper limb massage first), 36.7 (control first) SD: unknown Postnatal age: Within a week of birth Sex: Males: 42 Females: 38 Country: Hong Kong
Interventions	Total number of intervention groups: 1 Control group description: the control conditions were usual care but no massage, the infant simply being positioned with a quilt covering both upper limbs (for blinding) and the hand held to facilitate venipuncture. The physicians then performed the same method of venipuncture to the control group. Total number of participants randomised to each of the groups: 80 Specific intervention: upper limb massage Intervention details: 5 minutes before the venipuncture, the intervener applied massage to both upper limbs for 2 minutes. Following the massage, the physicians performed venipuncture. Massage was done by NICU nurse and it consisted of upper limb stroking, milking, gentle squeezing, and rolling. Integrity of Intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: 30 seconds post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: upper limb massage may be effective in decreasing infants' venipuncture pain perception Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	High risk	Using a list of random numbers is considered an open random allocation schedule and according to Cochrane, this is high risk.
3a) Performance bias (blinding of personnel)	High risk	Personnel not blinded to study condition.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were blinded to study condition.
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data; explained reasons for attrition from first venipuncture to second one. Final sample sizes for follow up were relatively equivalent.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain has a rating of high then the overall rating is high.

Chirico 2017.

Study characteristics
Methods	Study design: between‐group Total study duration: 24 months
Participants	Total number: 40 Setting: hospital (NICU) Diagnostic criteria: stable preterm neonates Corrected age: Mean: 33.4 (maternal voice), 33.3 (control), 33.0 (overall) SD: 2.0 (maternal voice), 1.9 (control) Sex: Males: not reported Females: not reported Country: Italy
Interventions	Total number of intervention groups: 1 Control group description: infants belonging to the control group were not exposed to maternal voice. Both groups received routine care, namely environmental and nonpharmacologic measures, according to Italian guidelines for procedural pain. Total number of participants randomised to each of the groups: Maternal voice: 20 Control: 20 Specific intervention: maternal voice Intervention details: a time was arranged with the mother to obtain a recording of her voice, either singing lullabies or saying nursery rhymes, according to her preference. The mother’s voice was recorded in agreement with the American Academy of Pediatrics guidelines that recommend the following safe sound levels within the NICU: an hourly personal equivalent noise exposure (Leq) of 50 decibels (dB); an hourly level exceeded for 10% of the time (L10) of 55 dB and a 1‐second duration maximum sound level (Lmax) of 70 dB. The volume of the recording was then normalised by a music editing software programme to avoid excessive peaks during the audio play‐back. Infants underwent 10‐minute‐long sessions of exposure to maternal voice, twice a day for 3 consecutive days. Subsequently, when the heel lance procedure was performed for clinical purposes, the maternal voice was played to infants belonging to the intervention group from 10 minutes before starting the heel lance procedure until 20 minutes after the end of the procedure. Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: 30 seconds post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: using recorded maternal voices to limit pain in preterm infants undergoing heel lance procedures appeared safe and effective Study funding source: this study did not receive any specific funding Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Authors do not sufficiently describe the method used to generate the randomisation.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Unable to blind personnel due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors were not blinded to study condition.
4) Attrition bias (incomplete outcome data)	Unclear risk	States that 40 infants were randomised (20 in each group), but no sample sizes are reported in the stats analysis so unclear whether they had data on all 40. No flow chart provided to explain any dropout or final N's. It was unclear how missing participants or data were accounted for.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall assessment is high risk.

Chou 2012.

Study characteristics
Methods	Study design: between‐group Total study duration: not mentioned
Participants	Total number: 149 Setting: hospital Diagnostic criteria: healthy neonate Postnatal age:  48 hours  Mean: not mentioned  SD: not mentioned  Sex: not mentioned  Country: Taiwan
Interventions	Total number of intervention groups: 1 Control group description: no pharmacological or non‐pharmacological interventions were applied Total number of participants randomised to each of the groups: Multisensory bundle: 88 Control: 99 Specific intervention #1: multisensory bundle Intervention details #1: 10 minutes before giving the vaccine, music was played to the infants. Infants’ breathing rate and heart rate were measured under normal situation, while taking the vaccine and 1 minute after vaccination. After vaccination, nurses soothed infants’ back for 3 minutes. Integrity of intervention #1: good
Outcomes	Pain outcomes, scale limits, and time points when measured NFCS Time points: first 30 seconds post‐procedure  Upper limit: 9 Lower limit: 0 Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors: behavioural interventions represent a simple, practical, and effective approach to alleviating neonate vaccination pain. Their proper application to improve neonatal care quality is recommended.  Study funding source: Medical Education Committee of Jen‐Ai Hospital Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	Unclear risk	It is not clear whether the personnel were blinded to the intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors unable to be blinded due to nature of intervention.
4) Attrition bias (incomplete outcome data)	Low risk	The number of enrolled participants in each group and the number reported in outcome tables match.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain is "high", the overall judgement is high.

Chuang 2019.

Study characteristics
Methods	Study design: cross‐over Total study duration: 9 months
Participants	Total number: 14 Setting: NICU and infant ward at a university medical centre Diagnostic criteria: preterm Gestational age: Mean: 26.8 SD: unknown Sex: Males: 6 Females: 8 Country: Taiwan
Interventions	Total number of intervention groups: 1 Control group description: standard care consists of partial developmental care such as positioning, containment, and oxygen supplementation as the routine care at the study site Total number of participants randomised to each of the groups: 14 Specific intervention: developmental care Intervention details: developmental care bundle consisted of environmental modifications, positioning and containment, oxygen supplementation, interaction and approach and cue‐based and individual care Integrity of Intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: 30 seconds post‐procedure and 1 minute post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: a bundled developmental care intervention significantly reduced pain following the procedure Study funding source: study was supported by the Ministry of Science and Technology in Taiwan Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Study personnel were not blinded to study condition.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors were not blinded because researcher who provided the intervention also assessed the outcome.
4) Attrition bias (incomplete outcome data)	Unclear risk	Flow chart does not include # of participants included in analysis; cannot find final sample sizes for the PIPP‐R anywhere; although not an outcome of interest, the sample sizes for recovery time were smaller than 14 (see abstract), but authors do not explain why some participants were not included in the analysis. It was unclear how missing participants or data were accounted for.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, because one domain received a rating of "high risk", the overall rating is high.

Cohen 2002.

Study characteristics
Methods	Study design: between‐group Total study duration: 1 year
Participants	Total number: 90 Setting: health department Diagnostic criteria: healthy infants Age: M = 12 months, SD = 8.6 months Sex: 44 male, 46 female Country: United States
Interventions	Total number of intervention groups: 2 Control group description: typical care, interact normally with infant but without movie or toy distraction Total number of participants randomised to each of the groups: Distraction: 49 Control: 41 Specific intervention: distraction Intervention details: teletubbies movie or toys, or both Integrity of intervention: poor
Outcomes	Pain outcomes, scale limits, and time points when measured MBPS Time points: 4 x 10‐second phases: baseline (20 seconds until 10 seconds prior to injection), anticipatory (10 seconds prior to injection until injection), injection (injection until 10 seconds later), and recovery (20 seconds following the final injection until 10 seconds later) Upper limit: 3.33 Lower limit: 0 High score = more pain Continuous VAS ‐ Parent During and 3 minutes after immunisation Upper limit: 100 Lower limit: 0 High score = more pain Continuous VAS ‐ Nurse During and 3 minutes after immunisation Upper limit: 100 Lower limit: 0 Higher score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: nurse‐directed distraction works but not sure if it is the toy, nurse, parent or combination Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Randomisation is mentioned but there is not an adequate description of the methods used.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Unclear risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	Low risk	Coders were blinded to study hypotheses.
4) Attrition bias (incomplete outcome data)	High risk	There is no documented evidence or insufficient evidence reported of how attrition was dealt with.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score on Cochrane standard: low risk (2/7), high risk (1/7), unclear risk (4/7)

Cohen 2006.

Study characteristics
Methods	Study design: between‐group Total study duration: 5 months (2 months in 2002, 3 months in 2003)
Participants	Total number: 136 Setting: health care facility Diagnostic criteria: any English‐speaking families with an infant between the ages of 1 and 24 months present for routine checkups and vaccinations Age: M = 7.6 months, SD = 5.0 months Sex: 80 girls and 55 boys Country: United States
Interventions	Total number of intervention groups: 1 Control group description: parent and nurse were encouraged to interact with the infant in their usual manner, no movie or toy distraction provided Total number of participants randomised to each of your groups: approximately 63 per group (total = 136) Specific intervention: distraction ‐ video Intervention details: prior to data collection, nurses engaged in brief intervention training and parents were briefly instructed in distraction techniques. During the immunisation, a DVD movie (choice between 2) played on a hand‐held DVD player 6 inches from the child. Parents were instructed that they could redirect attention to DVD. Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured Visual analogue scale (VAS) Time points: before, during, and after immunisation (spanning from 3 minutes prior to when the nurse began cleaning the skin for the injection, until 2 minutes after the needle was removed) Upper limit: 100 Lower limit: 0 High score = more pain Continuous MAISD (infant distress, crying, screaming or flailing) Coded for 5‐second periods up to 2 minutes post needle Upper limit: 1 Lower limit: 0 High score = higher pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: distraction is effective in reducing infants’ behavioural distress, in both the anticipatory and recovery phases of the immunisation Study funding source: this research was supported by a grant awarded by the Mayday Fund to Dr. Lindsey Cohen Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	High risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	High risk	Coders were not blinded.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score on Cochrane standard: low risk (4/7), high risk (2/7), unclear risk (1/7).

Comaru 2009.

Study characteristics
Methods	Study design: cross‐over Total study duration: April to October 2004
Participants	Total number: 47 Setting: NICU Diagnostic criteria: preterm Age: Minimum: not reported Maximum: 35 weeks Mean: 32 weeks SD: 2 weeks Sex: Males: not reported Females: not reported Country: Brazil Co‐morbidity: respiratory distress syndrome was diagnosed in approximately 35% of the studied babies Socio‐demographics: not reported Ethnicity: not reported
Interventions	Total number of intervention groups: 1 Control group description: babies' diaper changed in the incubator Total number of participants randomised to each of the groups: 47 Specific intervention: nesting Intervention details: slightly flexed posture, side lying, limbs directed to midline: head, back, links feet with folded up towel Integrity of intervention: not needed
Outcomes	Pain outcomes, scale limits, and time points when measured: NFCS Time points: not mentioned  Upper limit: 1 Lower limit: 0 Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: nesting works in diminishing distress during diaper changes Study funding source: this work was financially supported by the Research Support Fund (FIPE) of the Hospital de Clinicas in Porto Alegre and by a CNPq Scholarship Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Unclear risk	None specified.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (0/7), unclear risk (3/7)

Corbo 2000.

Study characteristics
Methods	Study design: cross‐over Total study duration: not reported
Participants	Total number: 26 Setting: NICU of the Division of Neonatology (University of Naples) Diagnostic criteria: neonates without severe complications Age: Minimum: 2 days Maximum: 15 days Mean: 4.9 days SD: 3.3 days Sex: Males: 16 Females: 10 Country: Italy
Interventions	Total number of intervention groups: 1 Control group description: infant in the supine position with their head towards the examiner, in incubator/crib/room, no pacifier Total number of participants randomised to each of the groups: 26 Specific intervention: non‐nutritive sucking (pacifier) Intervention details: the pacifier was placed in the infant's mouth 3 minutes pre‐procedure and for 3 minutes post‐procedure Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: Brazelton Neonatal Behavioral Assessment Scale (BNBAS) State 6 Time points: 3 minutes post‐procedure Upper limit: 180 seconds Lower limit: 0 seconds Higher score = longer time spent in distressed state Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: non‐nutritive sucking can reduce the period of behavioural distress in newborn, premature infants following heel stick blood sampling Funding sources: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported but this did not involve an independent person.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Unclear risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	High risk	Coders were not blinded to the condition.
4) Attrition bias (incomplete outcome data)	High risk	There is no documented evidence or insufficient evidence reported of how attrition was dealt with.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (3/7), high risk (1/7), unclear risk (3/7).

Corff 1995.

Study characteristics
Methods	Study design: cross‐over Study duration: not reported
Participants	Total number: 30 Setting: NICU Diagnostic criteria: less than 22 days old, and no anomalies Age: Minimum: 25 weeks Maximum: 35 weeks Mean: 30 weeks SD: unknown Sex: Males: not reported Females: not reported Country: USA
Interventions	Total number of intervention groups: 1 Control group description: normal nursery routine Total number of participants randomised to each of the groups: 30 Specific intervention: facilitated tucking Intervention details: facilitated tucking was supplied during and after the heel stick by one consistent neonatal nurse Integrity of intervention: poor (not reported)
Outcomes	Pain outcomes, scale limits, and time points when measured: Total crying time Time points: not mentioned  Upper limit: 100% Lower limit: 0 % Longer cry = more distress Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: facilitated tucking works Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	High risk	Randomisation is mentioned but there is not an adequate description of the methods used.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Unclear risk	None specified.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (2/7), high risk (2/7), unclear risk (3/7).

Cramer‐Berness 2005a.

Study characteristics
Methods	Study design: between‐group Total study duration: 12 months
Participants	Total number: 123 Setting: health care clinic in upstate New York Diagnostic criteria: healthy infants between 2 months and 24 months of age who were receiving routine immunisations Age: Minimum: 2 months Maximum: 24 months Mean: 8.7 months SD: 5.9 months Sex: Males: 60 Females: 63 Country: USA Co‐morbidity: none Socio‐demographics: never married = 59%; married = 28.7%; separated = 9%; divorced = 3.3% Less than USD $10,000 = 32.5%; $10,000 to $20,000 = 29.8%; $20,001 to $30,000 = 24.6%; greater than $30,001 = 13.2%; did not answer = 7.3% Ethnicity: Caucasian parents = 63.9% Caucasian infants = 51.3%; Hispanic parents = 13.4% Hispanic infants = 7.6%; African American parents = 12.6% African American infants = 15.1%; other parents = 10.1% other infants = 8.4%
Interventions	Total number of intervention groups: 2 Control group description: typical care ‐ parents did not receive prompting to engage in usual behaviours or distraction Total number of participants randomised to each of the groups: typical care = 41, supportive care = 42, distraction = 40 (total = 123) Specific intervention #1: supportive care Intervention details: parents were asked what they normally do to reduce infant distress. They were encouraged by the researcher to engage in these techniques. Integrity of intervention: satisfactory Specific intervention #2: distraction Intervention details: parents were encouraged to use a distraction toy and to direct infant's attention towards the toy (e.g. "Look at this!") Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured: MBPS Time points: 5 seconds pre‐needle, 5 seconds during needle, 5 seconds after needle Upper limit: 3 Lower limit: 0 High score = more distress Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: brief prompting for parents to use their normal coping strategies may be more effective than training in helping infants recover more quickly from routine procedure pain Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported but this did not involve an independent person.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Unclear risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	High risk	Coders were not blinded to intervention.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (1/7), unclear risk (2/7).

Cramer‐Berness 2005b.

Study characteristics
Methods	Study design: between‐group Total study duration: 14 months
Participants	Total number: 117 Setting: paediatrician's office Diagnostic criteria: healthy infants Age: Minimum: 2 months Maximum: 24 months Mean: 8.6 months SD: 5.8 months Sex: Males: 63 Females: 54 Country: USA Socio‐demographics: Annual household income: Less than USD $10,000: 29.1% $10,000 to $20,000: 33.3% $20,001 to $30,000: 10.3% $30,001 to $50,000: 10.3% $50,001 and higher: 6.8% Ethnicity: Caucasian parents: 64.1%; Caucasian infants: 55.6% Hispanic parents: 8.5%; Hispanic infants: 6.0% African American parents: 14.5%; African American infants: 12.0% Asian parents: 1.7%; Asian infants: 1.7% Native American parents: 2.6%; Native American infants: 1.7% Multiracial parents: 7.7%; multiracial infants: 23.1%
Interventions	Total number of intervention groups: 2 Control group description: typical care Total number of participants randomised to each of the groups: Audiovisual distraction: 41 Tactile distraction: 38 Control: 38 Specific intervention #1: audiovisual distraction Intervention details: parents were instructed to encourage infants to focus on toys throughout immunisation process by verbal statements such as "look at this!" and engaging behaviours, such as playing music with the toy Integrity of intervention: good Specific intervention #2: tactile distraction Intervention details: parents were instructed to tickle their infants during the immunisation and alternate where they tickled with each immunisation Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured: MBPS Time points: 5 seconds post‐needle Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: distraction (auditory or tactile) does not reduce pain in healthy infants aged 2 to 24 months during immunisation Miscellaneous comments by review authors: if more than one needle was given they averaged out the pain score between needles to account for 1 to 4 needles. Therefore, there was a separate score for "during needle" phase. Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	A convincing effort to reduce bias in outcome measurement is reported.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (0/7), unclear risk (2/7).

Curtis 2007.

Study characteristics
Methods	Study design: between‐group Total study duration: 6 months (2 x 3‐month periods)
Participants	Total number: 41 Setting: hospital Diagnostic criteria: all infants up to 6 months corrected age that required venipuncture; participants required to have had nothing by mouth for 5 minutes prior to venipuncture; could not be critically ill; no fructose intolerance; no EMLA at the site of venipuncture Age: Placebo M = 64 days (55 SD); pacifier and placebo M = 38 days (39 SD) Sex: Placebo: males = 13, females = 6; pacifier and placebo: males = 9, females = 13 Country: Canada
Interventions	Total number of intervention groups: 1 (placebo, and placebo and pacifier) Control group description: placebo group received 2 mL of sterile water, administered via syringe over tongue at 2 minutes prior to venipuncture Total number of participants randomised to each of the groups: Placebo: 19 Pacifier and placebo: 22 Specific intervention: placebo and pacifier Intervention details: timers used to co‐ordinate events; solution administered by research nurse to anterior aspect of tongue over 30 seconds via syringe and pacifier inserted orally (if deemed necessary) at 2 minutes post intervention. Venipuncture was performed by nurse. Parents interacted with voice and touch as per usual. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured FLACC Time points: before and after procedure (30 to 60 seconds post‐intervention) Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse reactions: there was one episode of vomiting which occurred in a total of 3 children
Notes	Key conclusions of study authors: pacifiers are effective analgesics Study funding source: this research was supported by the paediatric residents training committee research award to Dr. Curtis at the Department of Pediatrics, Stollery Children's Hospital Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	A convincing effort to reduce bias in outcomes measurement is reported.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Low risk	There was no selective outcome reporting.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score on Cochrane standard: low risk (6/7), high risk (0/7), unclear risk (1/7).

Dabas 2019.

Study characteristics
Methods	Study design: between‐group Total study duration (months): 2
Participants	Total number: 100 Setting: immunisation clinic Diagnostic criteria: healthy infants Postnatal age (weeks): 10 and 14  Mean: not mentioned  SD: not mentioned Sex: Males: not mentioned Females: not mentioned Country: India
Interventions	Total number of intervention groups: 4 Control group description: infants in the control group received immunisation without any distraction and only routine care was maintained. Total number of participants randomised to each of the groups: Electronic toy: 25 Key toy: 25 Simple toy: 25 Control: 25   Specific intervention #1: Electronic toy Intervention details #1: Infants in experimental groups were distracted by electronic toys during immunisation procedure. The distraction technique given by researcher started 30 seconds before immunisation and it lasted until 15 seconds after the immunisation of each infant. Integrity of intervention #1: poor   Specific intervention #2: Key toy Intervention details #2: Infants in experimental groups were distracted by key toys during immunisation procedure. The distraction technique given by researcher started 30 seconds before immunisation and it lasted until 15 seconds after the immunisation of each infant. Integrity of intervention #2: poor   Specific intervention #3:  Simple toy Intervention details #3:  Infants in experimental groups were distracted by simple toys during immunisation procedure. The distraction technique given by researcher started 30 seconds before immunisation and it lasted until 15 seconds after the immunisation of each infant. Integrity of intervention #3: poor
Outcomes	Pain outcomes, scale limits, and time points when measured FLACC Time points: during and post‐immunisation Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors: distraction technique was effective in reducing the pain intensity during immunisation among infants. The mean pain score of electronic toy group was lower as compared to key toy group, simple toy group, and control group. Study funding source: not reported   Author COI statement: there are no conflicts of interest in this study
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel unable to be blinded due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors unable to be blinded due to nature of intervention.
4) Attrition bias (incomplete outcome data)	Low risk	The number of enrolled participants in each group and the number reported in outcome tables match.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain is "high", the overall judgement is high.

Davari 2018.

Study characteristics
Methods	Study design: cross‐over Total study duration: not reported
Participants	Total number: 40 Setting: hospital (NICU) Diagnostic criteria: stable preterm neonates Age: 62.5% were between 32 to 36 weeks gestational age; 40% were 1 to 3 days old post‐natal age Sex: Males: 23 Females: 17 Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: in the routine position, the infant was put in supine position Total number of participants randomised to each of the groups: 40 Specific intervention: facilitated tucking Intervention details: during facilitated tucking position, a nurse placed the infant in the position 2 minutes before blood sampling. To perform this positioning, the nurse warmed her hands and the position was maintained for 1 minute after the blood sampling. Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure and 3 minutes after procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: overall findings showed that comparing neonates in the 2 positions, there was no significant difference in their pain intensity Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Authors do not sufficiently describe the method used to generate the randomisation.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Given that the infant is being physically manipulated differently in the two interventions, it is not possible to blind study personnel to study condition.
3b) Detection bias (blinding of outcome assessors)	High risk	While it was mentioned that the researcher and a skilled nurse completed reliability training, it is not entirely clear who completed the PIPP assessments and there is no mention of the assessors being blinded to the condition. Additionally, given that the PIPP was completed during blood sampling (and not afterward by video), blinding would likely not have been possible.
4) Attrition bias (incomplete outcome data)	Low risk	40 infants were randomised and data on 40 infants were analysed/reported.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, because at least one domain was high risk the overall assessment is high risk.

Dehghani 2019.

Study characteristics
Methods	Study design: between‐group Total study duration: 4 months
Participants	Total number: 66 Setting: hospital Diagnostic criteria: preterm Age (days): Mean: 8.5 SD: 4.06   Sex: Males: 54 Females: 45 *numbers only for the 2 groups of interest not reported, overall numbers are provided for sex Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: no intervention Total number of participants randomised to each of the groups: Yakson touch: 33 Control: 33 Specific intervention: Yakson touch Intervention details: researcher calmed both arms and the muscles of both shoulders for 1 minute and took a deep breath to have energy at her palm. Yakson touch takes 5 minutes, including 1 minute of hand rest, 3 minutes of gentle petting, and 1 minute of opening hand. The researcher is continuously touching the infant during touching the hand and all the fingers. The researcher lifts the infant in a way that her left hand is placed under the infant and her right hand is placed on the infant’s abdomen. The infant’s top of abdomen is touched clockwise in a circle with a diameter of about 4 cm every 5 seconds. Phlebotomy was performed 1 minute after the end of the touch. Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: the infants’ responses to pain were measured and recorded 1 minute before the intervention and immediately after phlebotomy for 1 minute Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: Yakson touch can reduce phlebotomy pain in preterm infants hospitalised in intensive care unit Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	Unclear risk	It is unclear whether the nurse performing the phlebotomy was in the room when the intervention was administered.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were stated to be blinded to group allocation.
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data observed.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, given that the rating of at least one domain was unclear, the overall rating was unclear.

Dezhdar 2016.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 57 Setting: hospital (NICU) Diagnostic criteria: preterm neonates Gestational age: Mean: 32.6 (swaddling), 33.2 (control) SD: 2.5 (swaddling), 2.4 (control) Sex: Males: 30 Females: 27 Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: not described Total number of participants randomised to each of the groups: Swaddling: 28 Control: 29 Specific intervention: swaddling Intervention details: the bare neonate was wrapped in a thin sheet with a clean diaper by the researcher, and using a stopwatch, 10 minutes after the swaddling began, the neonate’s hand was taken from the swaddle and blood sampling was performed by the same skilled nurse. After a dressing was placed on the blood sampling site, the neonate’s hand was placed back in the swaddle and the condition was kept for 2 minutes after blood sampling. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure and 1 minute after procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: the results of this study indicate that there is no meaningful difference between swaddling and KMC on physiological indexes and pain in neonates. Therefore, the swaddling method may be a good substitute for KMC. Study funding source: this study was supported by the deputy of research in Bushehr University of Medical Sciences Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Given that the neonates were physically manipulated differently in each condition, it was not possible for study personnel to be blinded to condition.
3b) Detection bias (blinding of outcome assessors)	Low risk	The neonate's face was the only part of the body video recorded (enabling blinding to condition by assessor), and the outcome assessor was blind to study condition.
4) Attrition bias (incomplete outcome data)	Low risk	While 8 participants dropped out following randomisation, the authors outline the reasons why. The reasons for loss to follow‐up were consistent across conditions and unlikely to be related to true outcome.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that there is a rating of high risk for one of the domains, the overall rating is high.

Dur 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: 12 months
Participants	Total number: 90 Setting: hospital (NICU) Diagnostic criteria: preterm infants hospitalised in the NICU Gestational age: Mean: 33.6 (Yakson touch + pacifier), 33.4 (gentle human touch + pacifier), 33.3 (control), overall mean = 33.4 SD: 1.4 (Yakson touch + pacifier), 1.6 (gentle human touch + pacifier), 2.2 (control), overall SD = 1.7   Sex: Males: 17 (Yakson touch + pacifier), 16 (gentle human touch + pacifier), 15 (control), total = 48 Females: 13 (Yakson touch + pacifier), 14 (gentle human touch + pacifier), 15 (control), total = 42 Country: Turkey
Interventions	Total number of intervention groups: 2 Control group description: all infants in the control, Yakson, and gentle human touch (GHT) groups were given pacifiers as the routine practice of the clinic before and during painful procedures. A pacifier was given to all infants 1 minute before heel lancing. Total number of participants randomised to each of the groups: Yakson touch + pacifier: 30 Gentle human touch + pacifier: 30 Control: 30 Specific intervention: Yakson touch + pacifier, gentle human touch + pacifier Intervention details: Yakson touch + pacifier: the practitioner relaxed both arms and muscles of both shoulders for 1 minute and took a deep breath to concentrate on the Ki energy in his palm. The Yakson method continued for 15 minutes with steady touch (5 minutes), compassionate caressing (5 minutes), and repetition of steady touch (5 minutes). In the Yakson touch method, the palms of the practitioner and all fingers were kept in close contact so that the infants did not feel pressure. Steady touch (5 minutes): One hand rested on the chest and abdomen of the preterm infant, while the other hand supported the back and hip of the preterm infant Compassionate Caress (5 minutes): In the same hand position, the practitioner repeated caressing and resting for 5 minutes: Caress (one minute), rest (30 seconds), caress (one minute), rest (30 seconds), and caress (2 minutes). The practitioner caressed the infants’ chest and belly with a 1 cm diameter clockwise circular movement every 10 seconds. Steady touch (5 minutes). The application started 5 minutes before heel lancing and continued until the completion of the procedure. Gentle Human Touch + Pacifier: While the practitioner placed one hand on the crown of the preterm infant on the eyebrow line with the fingertip touch for 15 minutes, the other hand was placed on the lower abdomen covering the waist and hip of the infant. The application started 5 minutes before heel lancing and continued until the completion of the procedure. They were given pacifiers as the routine practice of the clinic before and during painful procedures. A pacifier was given to all infants 1 minute before heel lancing. Integrity of intervention: Yakson touch + pacifier = good, gentle human touch + pacifier = good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: pain scores were recorded before the intervention. Pain scoring started at the moment the infant’s heel was punctured and terminated after collection of the necessary amount of blood. 5 minutes after the procedure, the infant’s NIPS score and physiologic parameters were evaluated and recorded. Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: tactile techniques (Yakson and GHT) performed during heel lancing process decreased pain scores and heart rates in preterm infants, but they did not affect oxygen saturation values. It can thus be deduced that the frequent use of Yakson and GHT in painful procedures such as heel lancing in preterm infants is effective in pain management and the regulation of physiologic parameters. Study funding source: no funding Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel (i.e. nurse) were not blinded to group allocation.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors were the same as those performing procedures and were therefore unblinded to group allocation.
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data observed.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Efendi 2018.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 30 Setting: special care unit and NICU Diagnostic criteria: preterm neonates (between 26 to 36 weeks gestation) Gestational age: Mean: not reported SD: not reported Sex: Males: not reported Females: not reported Country: Indonesia
Interventions	Total number of intervention groups: 1 Control group description: routine care Total number of participants randomised to each of the groups: Pacifier + swaddling: 15 Control: 15 Specific intervention: pacifier + swaddling Intervention details: the intervention (pacifier and swaddling) was applied 3 minutes before the invasive procedure started Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: 3 minutes after procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: providing pacifier and swaddling can impede the increase of premature infants’ pain score and heart rate during an invasive procedure, therefore it can be implemented as an alternative to pain management in premature infants Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Authors do not sufficiently describe the method used to generate the randomisation.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	While it would likely not be possible to fully blind study personnel given the nature of the study condition, there is no mention by the authors of any study personnel being blinded to study condition.
3b) Detection bias (blinding of outcome assessors)	High risk	No mention of blinding the outcome assessors to study condition.
4) Attrition bias (incomplete outcome data)	Low risk	While no attrition or exclusions from analyses mentioned, they recruited 30 infants and reported data on 30 infants.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that at least one domain was given the rating of "high risk", the overall risk is high.

Elserafy 2009.

Study characteristics
Methods	Study design: cross‐over Study duration: January 2005 and May 2007
Participants	Total number: 36 Setting: NICU tertiary care Diagnostic criteria: exclusion criteria were 1) exposure to maternal sedation; 2) occurrence of any procedure performed within 24 hours in preterm infants whose mothers had had general anaesthesia during delivery; 3) the presence of major neurologic abnormalities; 4) Apgar scores at 5 minutes of greater than 5; 5) presence of necrotising intestinal colitis; 6) nothing by mouth status for any reason; 7) being preterm with hyperglycaemia Age: Minimum: 27 weeks Maximum: 36 weeks Mean: 32.4 weeks SD: 2.9 Sex: Males: not reported Females: not reported Country: Saudi Arabia
Interventions	Total number of intervention groups: 3 Control group description: standardised care Total number of participants randomised to each group: 36 Specific intervention #1: water pacifier Intervention details: infants were given pacifiers coated in 0.5 mL of water 2 minutes prior Integrity of intervention: good Specific intervention #2: pacifier alone Intervention details: standard nipple stuffed with gauze square for resistance, held in infant's mouth for 2 minutes prior Integrity of intervention: good Specific intervention #3: sterile water Intervention details: 0.5 mL of sterile water without pacifier Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: PIPP Time points: 15 seconds after the venipuncture Upper limit: 3 Lower limit: 0 Higher score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: water pacifier, pacifier alone, and sterile water were not as effective as sucrose and pacifier Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Unclear risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (0/7), unclear risk (3/7).

Erdogan 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: 7 months
Participants	Total number: 60 Setting: hospital (PICU) Diagnostic criteria: the children [on the unit] were admitted to the paediatric ICUs mainly due to respiratory disorders, need for mechanical ventilation or dialysis or life‐threatening illnesses. The unit patient population had a wide range of diagnoses, ranging from cerebral palsy to suicide, ketoacidosis and acute and chronic renal failure. Age: Intervention: mean = 2.0 years (0.8) Control: mean = 2.0 years (0.2)   Sex: Intervention: 11 females, 19 males Control: 13 females, 17 males Overall: 24 females, 36 males Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: standard procedure Total number of participants randomised to each of the groups: Intervention: 30 Control: 30 Specific intervention: maternal voice Intervention details: the mothers of the children in the experimental group were given a voice recorder and explained that they could tell a story, sing a song or talk to their children by recording the same into the voice recorder, which their children would listen to during painful procedures. It was clearly explained to the parents that the goal with this practice was to distract the children from the process. The child would listen with the help of a headset to cover 5 minutes before and after the procedure to be applied. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured FLACC Time points: pain levels were evaluated at baseline and twice, during and after the procedure Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: pain levels and heart rates of the children in the experimental group during and after the painful procedures were much lower, and their oxygen saturations were higher than those of the children in the control group. Study funding source: the study received no funding Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	One of the limitations is the use of the urn method for randomisation.
2) Selection bias (allocation concealment)	Low risk	Adequately described.
3a) Performance bias (blinding of personnel)	High risk	Personnel would see headset on infant.
3b) Detection bias (blinding of outcome assessors)	High risk	There is a lack of blinding with only one nurse researcher reporting the pain level (FLACC).
4) Attrition bias (incomplete outcome data)	Low risk	60 participants were randomised and 60 participants were included in analyses.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall assessment is high risk.

Erkut 2017.

Study characteristics
Methods	Study design: between‐group Total study duration: 3 months
Participants	Total number: 74 Setting: hospital Diagnostic criteria: healthy full‐term neonates Gestational age: Percentages: 70.3% were 39 weeks (swaddling), 56.8% were 39 weeks (control) Postnatal age: Percentages: 54.1% were 24 hours (swaddling), 67.6% were 24 hours postnatal (control) Sex: Males: 55.9% (swaddling), 51.4% girls (control) Females: 54.1% (swaddling), 48.6% girls (control) Country: Indonesia
Interventions	Total number of intervention groups: 1 Control group description: for the control group, heel lance was performed with newborns in a "natural position". The natural position was the one taken by the newborns themselves while lying in a supine position on the procedure table. No comforting or relaxation interventions were used during the heel lance procedure. After the procedure these newborns were immediately comforted. Total number of participants randomised to each of the groups: Swaddling: 37 Control: 37 Specific intervention: swaddling Intervention details: for newborns in the experimental group, newborns were swaddled with legs in flexion and abduction without causing any movement restriction while lying in a supine position on the procedure table. During the heel lance procedure parents were not allowed to hold the newborns. After the procedure newborns were immediately comforted by their parents. Swaddling was performed one minute before heel lance and maintained for 3 minutes after the procedure. Swaddling was not too firm such that it would inhibit behavioural indicators (such as arm and leg movements) used in the pain assessment; neither was it too loose in order to protect the experiment’s integrity. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time Points: first 30 seconds post‐procedure and 3 minutes post‐procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: for this study sample, swaddling was an effective nonpharmacologic method to help reduce pain and crying in an effort to soothe newborns Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	While the researchers described the standardisation procedure across, personnel were not blinded (and likely could not have been) to study condition.
3b) Detection bias (blinding of outcome assessors)	High risk	Pain assessors (nurse and researcher) were not blinded to study condition when scoring the NIPS.
4) Attrition bias (incomplete outcome data)	Low risk	74 participants were randomised and data on 74 participants were reported/analysed.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Esfahani 2013.

Study characteristics
Methods	Study design: between‐group Total study duration: April 2011 to July 2011 (3 months)
Participants	Total number: 64 Setting: Navabsafabi Health Care Center Diagnostic criteria: healthy neonates Age Minimum: 6 months Maximum: 12 months Mean: N/A SD: N/A Sex Males: N/A Females: N/A Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: vaccination was conducted in the conventional method with no intervention Total number of participants randomised to each of the groups: Intervention: 32 Control: 32 Intervention descriptions: massage therapy: researcher massaged the first knuckle of the middle or ring finger of the infants’ palm or sole of the injection side for 60 seconds and finally vaccinated the participant. The massage point was corresponding with the injection‐related leg in the body through sujok correspondence system; the points are on a miniature map of the body on the palms and soles. Integrity of intervention: poor
Outcomes	Pain outcomes, time points when measured, and scale limits: NIPS Time points: “During the procedure of vaccination” page 3, paragraph one under Materials and Methods – did not specify timing Range of possible scores: 0 (lowest pain score) to 7 (highest pain score); score of 3 shows existence of pain Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: massage therapy can be used as a replacement method for breastfeeding (less vaccination pain reported in the massage group compared to the control group); breastfeeding was more effective than massage therapy Study funding source: Isfahan University of Medical Sciences Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	“The type of vaccine to be investigated was randomly selected from 96 envelopes marked as A and B, which the researcher had already made by random number table (zero was ignored; numbers 1, 2, 3 were assigned to the massage group; 4, 5, 6 to the breast feeding group; and 7, 8, 9 to the control group. Envelopes A were assigned to six‐month‐old infants and envelopes B to 12‐month‐old infants”, page 3, paragraph 2 under Materials and Methods.
2) Selection bias (allocation concealment)	Low risk	Same as above.
3a) Performance bias (blinding of personnel)	High risk	Participants and personnel were not blinded because they would see which intervention the child was given.
3b) Detection bias (blinding of outcome assessors)	High risk	Coder were not blinded as they evaluated NIPS in the room and would have seen the intervention the infant was assigned to.
4) Attrition bias (incomplete outcome data)	Low risk	A total of 99 infants entered the study, 3 infants were left out due to restlessness and cry before the injection.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (2/7), unclear risk (0/7).

Fatollahzade 2020.

Study characteristics
Methods	Study design: cross‐over Total study duration: 19 months
Participants	Total number: 34 Setting: hospital (NICU) Diagnostic criteria: preterm neonates requiring endotracheal suctioning Gestational age: < 28 weeks = 6 infants, 28 to 31 weeks = 7 infants, 32 to 34 weeks = 21 infants Sex: Males = 19, females = 15 Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: the research assistant had 10 years of experience working in the neonatal intensive care unit, assisting the principal researcher in all stages of intervention and measurement, and was performing suctioning in all cases according to standard guidelines. All suctioning (both by the routine method and by gentle human touch) was performed by the same nurse. Total number of participants randomised to each of the groups: 34 Specific intervention: gentle human touch suction Intervention details: the neonate is placed in a lying position and gently flexed in a position similar to the position of a curved fetus naturally assumed in the womb as they near the term gestation. This flexible position has a more stabilising and relaxing effect on the premature infant. Gentle human touch (GHT) is provided by cupping one hand around the infant’s head while cupping the other hand around the infant’s bottom. This position allowed the infant to move but with limited extension and flailing of the extremities. Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: immediately after suction Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: results from this study showed that the pain due to suctioning procedure is considerably reduced by applying gentle human touch. Nurses can use this method as one of the non‐pharmacological methods of pain management. Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	There is not an adequate description of method for generation random allocation sequence.
2) Selection bias (allocation concealment)	High risk	A high risk method for dealing with allocation bias is described. This cross‐over trial used an open random allocation schedule whereby the first half of the sample received the procedure without the intervention and then with the intervention. This order was reversed for the second half of the sample.
3a) Performance bias (blinding of personnel)	High risk	Personnel (i.e. nurse) were not blinded to group allocation.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors were not blinded to group allocation.
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data observed.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Fearon 1997.

Study characteristics
Methods	Study design: cross‐over Total study duration: not reported
Participants	Total number: 15 Setting: NICU Diagnostic criteria: preterm Age: Minimum: 27 weeks Maximum: 36 weeks Mean: younger group (29.6 weeks) and older group (33.7 weeks) SD: younger group (0.9 weeks) and older group (2.2 weeks) Sex: Males: 9 Females: 6 Country: Canada Socio‐demographics: not collected Ethnicity: not collected
Interventions	Total number of intervention groups: 1 Control group description: typical care Total number of participants randomised to each group: 15 Specific intervention: swaddling Intervention details: placed infant in a supine position on a soft clean cloth. Infant's arms were crossed over in a relaxed position and tucked in on sides. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: NFCS Time points: pre‐needle, during needle, and post‐needle Upper limit: 10 Lower limit: 0 Higher score = more pain Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: pain behaviours reduced by swaddling Study funding source: this study was funded in part by an Ontario Ministry of Health Nursing Innovation Fund grant to JT, BSK, and DWM. BSK is supported by an Ontario Ministry of Health Career Scientist Award. Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported but this did not involve an independent person.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	A convincing effort to reduce bias in outcomes measurement is reported.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (0/7), unclear risk (2/7).

Fitri 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: 5 months
Participants	Total number: 72 Setting: hospital Diagnostic criteria: infants are 35 weeks; description would meet the preterm category Gestational age: Mean: 35.9 (sensory stimulation with sucrose), 36.1 (oral sucrose alone/control group) SD: 2.9 (sensory stimulation with sucrose), 1.8 (oral sucrose alone/control group)   Sex: Males: 19 (sensory stimulation with sucrose), 20 (oral sucrose alone/control group) Females: 17 (sensory stimulation with sucrose), 16 (oral sucrose alone/control group) Country: Indonesia
Interventions	Total number of intervention groups: 1 Control group description: for the control group (S24%), 30 seconds before and throughout the blood sampling procedure, sucrose 24% was administered Total number of participants randomised to each of the groups: Sensory stimulation with sucrose: 36 Oral sucrose alone/control group: 36 Specific intervention: sensory stimulation with sucrose Intervention details: 30 seconds before collecting blood, the nurses started talking to the baby, looked into the baby's eyes, gently massaged the cheeks and other areas of the baby's face, and administered sucrose 24% until the baby sucked rhythmically; then, the nurse in charge of obtaining the blood sample inserted the needle in the dorsal vein of one of the baby's hands, and smooth massage was carried out. The assessment of pain intensity was conducted by a different nurse from the time of needle puncture until the procedure was completed. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP‐R Time points: the assessment of pain intensity was conducted from the time of needle puncture until the procedure was completed Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: multisensory stimulation is more effective in reducing pain than unimodal (oral sucrose) analgesia Study funding source: this study was funded by the Ministry of Research, Technology and Higher Education of Republic of Indonesia Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel were not blinded to group allocation.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors were not blinded to group allocation.
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data observed.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Gao 2018.

Study characteristics
Methods	Study design: between‐group Total study duration: 7 months
Participants	Total number: 86 Setting: hospital (NICU) Diagnostic criteria: preterm neonates Gestational age: Mean: 31.9 (NNS), 31.7 (oral sucrose), 32 (NNS + sucrose), 31.3 (control) SD: 1.1 (NNS), 0.9 (oral sucrose), 0.8 (NNS + sucrose), 0.9 (control) Sex: Males: 52 Females: 34 Country: China
Interventions	Total number of intervention groups: 3 Control group description: the preterm infant received only routine comfort through gentle touch when he cried after the heel stick procedure. For ethical reasons, preterm infants in the routine care group were given gentle touch to alleviate procedural pain if they were crying. Total number of participants randomised to each of the groups: NNS: 22 Sucrose: 21 NNS + sucrose: 22 Control: 21 Specific intervention #1: NNS Intervention details: the preterm infant was given a standard silicone newborn pacifier to stimulate sucking 2 minutes before, and throughout the recovery phase of the heel stick Integrity of intervention #1: good Specific intervention #2: sucrose Intervention details #2: sucrose 20% (0.2 mL/kg) was administrated to the preterm infant’s mouth by 1 mL syringe without the needle in 2 minutes before the heel stick procedure Integrity of intervention #3: good Specific intervention #3: NNS + sucrose Intervention details #3: sucrose 20% (0.2 mL/kg) was administrated to the preterm infant’s mouth by 1 mL syringe without the needle in 2 minutes before the heel stick procedure, and then a standard silicone new‐born pacifier was given to stimulate sucking until the recovery phase of the heel stick Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure and 1 minute post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: 3 preterm infants vomited (1 in the combination group, 1 in the routine care group, 1 in the non‐nutritive sucking group); 2 preterm infants had abdominal distension (1 in the combination group, 1e in the routine care group)
Notes	Key conclusions of study authors: the combination of non‐nutritive sucking with sucrose provided better pain relief during repeated painful procedures than when non‐nutritive sucking or sucrose was used alone. The effect of non‐nutritive sucking was similar to that of sucrose on repeated procedural pain. Study funding source: Financial contributions of: National Natural Science Foundation of China, the Preponderant Discipline Project of Universities in Jiangsu Province, Nursing Science Open Fund of Nanjing University of Chinese Medicine, Top‐notch Academic Programs Project of Jiangsu Higher Education Institutions, and Project of nursing science in Nanjing University of Chinese Medicine Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported. Preterm infants were randomly allocated using a random computer‐generated table to one of the 4 groups.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Blinding was not possible given nature of intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	While outcome assessors were unaware of the purpose of the study, they were still not blinded to study condition.
4) Attrition bias (incomplete outcome data)	Low risk	Reasons for participant dropout were clearly explained, with similar reasons for missing data across groups.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall assessment is high risk.

Gormally 2001.

Study characteristics
Methods	Study design: randomised between groups Total study duration: not reported
Participants	Total number: 41 Setting: community hospital Diagnostic criteria: healthy infants born at term Age: Minimum: 37 weeks Maximum: not reported Mean: 39.3 SD: 1.3 Sex: Males: 22 Females: 19 Country: Canada Socio‐demographics: mothers in control group had a mean age of 29.9 years and 10.7 years of education; mothers in holding group had a mean age of 29.4 years and 11.4 years of education
Interventions	Total number of intervention groups: 1 Control group description: no holding and water taste Total number of participants randomised to each of the groups: Intervention: 20 Control: 21 Specific intervention: holding and water taste Intervention details: the infant is held in the arms of a female RA beginning 4 minutes pre heel stick, given sterile water and rested in the morning Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: NFCS (% of the time) 1 minute post‐heel lance, 2 minutes post‐heel lance Upper limit: 100 Lower limit: 0 Higher score = more pain Continuous Cry (% of the time) One, 2, and 3 minutes post‐heel lance Upper limit: 100 Lower limit: 0 Higher score = more time spent crying/more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: providing care‐giving context reduces pain in neonates Study funding source: this study was supported by the Medical Research Council of Canada (MT‐13633), the Allan Ross Fellowship and the Lewis Sessenwein Academic Award Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported but this did not involve an independent person.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Unclear risk	None specified.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (3/7), high risk (1/7), unclear risk (3/7).

Gray 2015.

Study characteristics
Methods	Study design: between‐group Total study duration: 2 months
Participants	Total number: 29 Setting: hospital Diagnostic criteria: healthy full‐term neonates Gestational age: Mean: 39 (overall) SD: 0.6 (overall) Sex: Males: 17 Females: 12 Country: USA
Interventions	Total number of intervention groups: 2 Total number of participants randomised to each of the groups: Sucrose: 15 Sucrose + warmer: 14 Specific intervention #1: sucrose Intervention details: infants in the sucrose alone group rested quietly in their bassinets, clothed in a diaper and shirt and unswaddled for the duration of the study. The protocol consisted of a baseline period (5 minutes), intervention (2 minutes), followed by the vaccination (10 seconds), and a recovery period (5 minutes). Infants in the sucrose plus warmer group were given 0.24 g of sucrose with the infant warmer increased to create a 0.5°C temperature gradient between the baby and the radiant warmth control temperature. Integrity of intervention #1: good Specific intervention #2: sucrose + warmer Intervention details #2: infants in the warmer plus sucrose group were placed under the Ohmeda Ohio Infant Warmer (ModelNo. 3000; GE Healthcare, Fairfield, CT), and their clothing was removed, except for a diaper. As a precaution against overheating or overcooling, infants were connected to the warmer’s servo control and temperature monitoring system at all times. The protocol consisted of a baseline period (5 minutes), intervention (2 minutes), followed by the vaccination (10 seconds), and a recovery period (5 minutes). During the 2‐minute intervention period, infants in the sucrose alone group were given 0.24 g of sucrose. After the vaccination, the radiant warmer was returned to the automatic or servo control setting. Integrity of intervention #2: good
Outcomes	Pain outcomes, scale limits, and time points when measured Duration of crying Time points: 4 minutes post‐procedure High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: the combination of sucrose and radiant warmth is an effective analgesic in newborns and reduces pain better than sucrose alone Study funding source: National Institute of Health/National Institute of Child Health and Human Development grants Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	It is unclear how exactly the researchers randomised participants. While they describe using sealed envelopes, it is not clear how the sequence was randomly generated.
2) Selection bias (allocation concealment)	Unclear risk	Authors note that the envelopes with intervention allocation were sealed, but did not mention they were sequentially numbered and opaque. There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	While it was likely not possible to blind study personnel to condition given the presence of the warmer in one of the conditions, authors did not mention any attempt to blind personnel.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were blinded to study condition.
4) Attrition bias (incomplete outcome data)	Unclear risk	Only those with "minimal recording artifact, sufficient for data analyses" were included in heart rate and RSA analyses ‐ not clear what is meant by "sufficient for data analyses" and whether this may be biased. Also, while the authors stated that 15 infants per group was the target sample size, they reported on 15 in one group and 14 in the other without an explanation of why. It was unclear how missing participants or data were accounted for.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain is designated as "high risk" then there is an overall rating of high risk.

Greenberg 2002.

Study characteristics
Methods	Study design: randomised between groups Total study duration: not reported
Participants	Total number: 42 Setting: moderate sized hospital, nursery Diagnostic criteria: birth weight between 2500 g and 4000 g; less than or equal to 72 hours old; no congenital abnormalities; full‐term Age: Minimum: age of birth Maximum: 72 hours Mean: 19.0 hours SD: 5.6 hours Sex: Males: 38 Females: 46 Country: USA Co‐morbidity: none Socio‐demographics: mothers were excluded from the study if there was a history of substance abuse, or chronic or infectious diseases during pregnancy Ethnicity: 73.8% white; 10.7% Hispanic; 4.8% African American; 10.7% other
Interventions	Total number of intervention groups: 3 (we will only focus on one for review) Control group description: routine care; no intervention offered during heel stick Total number of participants randomised to each of the groups:  Control = 21 Intervention = 21 Specific intervention: water pacifier on top of sucrose Intervention details: a Mini‐Mam 0 to 6‐month orthodontic pacifier moistened with water was gently held in the infant's mouth (without initiating any other contact) during 2 minutes pre‐procedure and 3 minutes post‐procedure Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured: Duration of cry Time points: from the start of the procedure to 25 minutes post‐procedure Upper limit: 1500 seconds (and higher) Lower limit: 0 seconds Higher score = longer cry duration, more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Funding sources: American Cancer Society; California Division Fellowship Key conclusions of the study authors: no significant differences between water pacifier and typical care group were found
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported but this did not involve an independent person.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	High risk	Investigator collected all data.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	High risk	There is no documented evidence or insufficient evidence reported of how attrition was dealt with.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (2/7), high risk (3/7), unclear risk (2/7).

Göl 2017.

Study characteristics
Methods	Study design: between‐group Total study duration: 5 months
Participants	Total number: 64 Setting: clinic (family health centre) Diagnostic criteria: healthy infants (4 to 6 months old) Postnatal age: Mean: 5 months (overall) SD: one month (overall) Sex: Males: 32 Females: 32 Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: the infants were seated on their mothers’ knees. Then DTaP/IPV/Hib was administered by using the conventional injection technique. Total number of participants randomised to each of the groups: Intervention: 32 Control: 32 Intervention descriptions: the infants were seated on their mothers’ knees. Next, 10‐second manual pressure was applied on the vaccine injection site. Then DTaP/IPV/Hib was administered in accordance with the conventional injection technique. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: first 30 seconds post‐procedure and 2 minutes post‐procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: manual pressure is effective and useful in relief of pain and reduction of crying time due to vaccine injections in 4‐ to 6‐month‐old infants Study funding source: this study was funded by Turkish Scientific and Technological Research Organization in the framework of a project numbered 214S387 and presented as a Doctoral dissertation at Ege University Health Sciences Institute Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	No mention of attempts at blinding to allocated interventions during the study.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcomes assessors were blinded to intervention allocation.
4) Attrition bias (incomplete outcome data)	Low risk	No missing data.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, an overall score of 'unclear' is given when one or more domain is determined to be unclear.

Harrington 2012.

Study characteristics
Methods	Study design: between‐group Total study duration: 6 months
Participants	Total number: 230 Setting: hospital Diagnostic criteria: Inclusion criteria: infants with a gestational age between 32 and 42 weeks at delivery and post natal age of < 20 weeks Exclusion criteria: acetaminophen or ibuprofen administration within 4 hours before immunisation, current neurological disorder, known genetic anomaly, moderate to severe illness with or without fever at the time of the immunisation, anaphylactic reaction to previous dose of vaccine, or infant was previously enrolled in the study at 2 months Age: 73% were 2 month olds Sex: Males: 124 Females: 106 Country: United States
Interventions	Total number of intervention groups: 3 Control group description: 2 mL of water 2 minutes before needle and comfort by parent or guardian after immunisation Total number of participants randomised to each of the groups: control = 56, sucrose = 58, physical = 58, physical and sucrose = 58 Specific intervention: sucrose group, physical group, physical and sucrose group Intervention details: sucrose group = 2 mL of 24% oral sucrose 2 minutes before immunisation and comfort by parent or guardian after immunisation. Physical group = 2 mL of water 2 minutes before immunisation and physical intervention using the 5 S’s by researcher after immunisation. Physical and sucrose group = 2 mL of 24% oral sucrose 2 minutes before immunisation and physical intervention using the 5 S’s by researcher after immunisation Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured The Modified Riley Pain Score Time points: 15‐second intervals for 2 minutes, then every 30 seconds up to 5 minutes post vaccination Upper limit: 9 Lower limit: 0 High score = more pain Continuous Crying Time points: 15‐second intervals for the first 2 minutes post last needle, then 30‐second intervals to a maximum of 5 minutes post last needle Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: physical intervention of the 5 S’s provided decreased pain scores on a validated pain scale and decreased crying time among 2 and 4 month old infants during routine immunisations. The use of the 5 S’s did not differ from the 5 S’s plus sucrose Study funding source: no external funding Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	“Infants of parents or guardians who consented were randomly assigned to 1 of 4 study groups by using presealed cards.” (p. 817)
2) Selection bias (allocation concealment)	Low risk	“Infants of parents or guardians who consented were randomly assigned to 1 of 4 study groups by using presealed cards. The cards were then selected by the nurses and only reviewed by the nurses to instruct them as to the group assignment” (p. 817)
3a) Performance bias (blinding of personnel)	High risk	Infants and parents not blinded to whether the physical intervention was used.
3b) Detection bias (blinding of outcome assessors)	High risk	Coders not blinded to whether infant received physical intervention.
4) Attrition bias (incomplete outcome data)	Low risk	There is no documented evidence or insufficient evidence reported of how attrition was dealt with.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score on Cochrane standard: low risk (5/7), high risk (2/7), unclear risk (0/7).

Hashemi 2016.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 67 Setting: hospital Diagnostic criteria: full‐term healthy neonates (postnatal age less than 3 days) Gestational age: Mean: 38.8 (swaddling), 38.5 (control) SD: 1.0 (swaddling), 1.1 (control) Postnatal age: Mean: 1.7 (swaddling), 1.7 (control) Sex: Males: 58.8% (swaddling), 60.6% (control) Females: 41.2% (swaddling), 39.4% (control) Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: no intervention Total number of participants randomised to each of the groups: Intervention: 34 Control: 33 Intervention descriptions: the swaddled group were swaddled a few minutes before vaccination and a few minutes later, while more than 45 minutes had passed from being breast fed (n = 34). The infants in the swaddled group were swaddled 2 minutes before and 2 minutes after vaccination. Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured: NFCS Time points: first 30 seconds post‐procedure Upper limit: 9 Lower limit: 0 Higher score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: regarding the significant impact of swaddling on the pain reduction of vaccination, it is recommended to take benefit of this safe and available non‐pharmacological method in order to relief pain during painful procedures Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Authors do not sufficiently describe the method used to generate the randomisation.
2) Selection bias (allocation concealment)	Unclear risk	Authors describe the use of sealed envelopes but to not mention sequentially ordered or opaque envelopes. There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Due to the differences in the intervention types (e.g. swaddled versus not swaddled), it was not possible to blind study personnel to intervention status.
3b) Detection bias (blinding of outcome assessors)	Low risk	Recording of face done in such a way that enables blind assessment of outcomes. Research assistant that scored NFCS was blinded to study condition.
4) Attrition bias (incomplete outcome data)	Low risk	Reasons for missing outcome data described and unlikely to be related to true outcome.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain has a rating of high then the overall rating is high.

Herrington 2007.

Study characteristics
Methods	Study design: cross‐over Total study duration: 4 months (October 2006 to January 2007)
Participants	Total number: 11 (21 enrolled initially; one withdrew, 2 were excluded, 6 discharged, one excluded due to video difficulty) Setting: NICU Diagnostic criteria: premature Age: Minimum: 27 weeks Maximum: 33 weeks Mean: 31 weeks SD: not reported Sex: Males: 6 Females: 5 Country: USA Ethnicity: 57.5% of the infants were African American; 40% Caucasian; 2.5% Hispanic origin
Interventions	Total number of intervention groups: 1 Control group description: no intervention, PI did not directly touch the infant. Both hands were loosely fisted and placed behind the infant in close approximation to the head and buttocks, but not in contact with the infant. Total number of participants randomised to each of the groups: 11 Specific intervention: gentle human touch Intervention details: gentle positioning support for the infant using warm human hands to contain the infant's body in a flexed position. PI placed both hands in isolette, placing the right hand behind the infant's head and shoulders in a cupped fashion. PI paused for 10 seconds to release tension. The position was held for the entire duration of the heel stick procedure and 2 minutes post heel stick procedure. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: Cry duration Time points: heel stick and 2 minutes post‐draw recovery Upper limit: (variable) Lower limit: 0 seconds Higher score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Funding sources: not reported Key conclusions of the study authors: gentle human touch does not reduce cry duration in moderately premature infants receiving the heel stick procedure Miscellaneous comments by review authors: study was double‐blind; participant information is not known
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	A convincing effort to reduce bias in outcomes measurement is reported.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (0/7), unclear risk (2/7).

Herrington 2014.

Study characteristics
Methods	Study design: repeated‐measures, cross‐over design Total study duration: not reported
Participants	Total number: 11 Setting: NICU, Sparrow Hospital Diagnostic criteria: premature Age: Minimum: 27 weeks Maximum: 34 weeks Mean: 31 weeks SD: 1.7 weeks Sex Males: 6 Females: 5 Country: USA
Interventions	Total number of intervention groups: 1 Control group description: standard care – infants also placed in a side‐lying position and supported with blanket rolls to create a “nest confinement” Total number of participants randomised to each of the groups: Intervention: 11 Control: 11 Intervention descriptions: gentle human touch was accomplished by placing the infant in a side‐lying position and gently flexing the infant into a tucked position similar to the flexed fetal position infants naturally assume in the womb as they near term gestation. GHT is provided by cupping one hand around the infant’s head while cupping the other hand around the infant’s bottom. This position allowed the infant to move but limited extension and flailing of the extremities Integrity of intervention: good
Outcomes	Pain outcomes, time points when measured, and scale limits: Audible cry Time points: immediately after needle, recovery not specified Range of possible scores: measured in seconds Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: gentle human touch reduces pain in premature infants undergoing heel stick Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	“Infants served as their own controls with random assignment to order of treatment” (p. 109, paragraph 1) ‐ this means that their own scores in the intervention and control group were compared to each other. No mention of how random assignment was generated.
2) Selection bias (allocation concealment)	Unclear risk	Did not address this outcome.
3a) Performance bias (blinding of personnel)	Low risk	“... phlebotomist at the bedside so that the phlebotomist was unaware of the experimental assignment of the infant” – p.110
3b) Detection bias (blinding of outcome assessors)	Unclear risk	Did not address this outcome.
4) Attrition bias (incomplete outcome data)	Low risk	Figure 3 page 111 outlines enrolment and exclusion of participants, repeated‐measures design so same for both conditions.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (0/7), unclear risk (3/7).

Hill 2005.

Study characteristics
Methods	Study design: cross‐over Total study duration: not reported
Participants	Total number: 12 Setting: NICU Diagnostic criteria: preterm Age: Minimum: 25 weeks Maximum: 34 weeks Mean: 30.9 weeks SD: 2.5 weeks Sex: Males: 6 Females: 6 Country: USA
Interventions	Total number of intervention groups: 1 Control group description: standard care Total number of participants randomised to each of the groups: 12 Specific intervention: facilitated tucking Intervention details: nurse provided care to the infant while the physical therapist supported the infant in a midline, tucked position Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: PIPP Time points: immediate, during procedure Upper limit: 21 Lower limit: 0 Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of the study authors: by incorporating facilitated tucking into routine care events, the stress levels of infants born preterm may be reduced Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported but this did not involve an independent person.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	A convincing effort to reduce bias in outcomes measurement is reported.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (0/7), unclear risk (2/7).

Hillgrove Stuart 2013.

Study characteristics
Methods	Study design: randomised between groups Total study duration: 11 months
Participants	Total number: 99 Setting: paediatrician's clinic Diagnostic criteria: healthy infants Age: Minimum: 12 weeks Maximum: 20 weeks Mean: 15.0 weeks SD: 2.9 weeks Sex: Males: 58 Females: 41 Country: Canada Socio‐demographics: predominantly moderate‐high socioeconomic class Ethnicity: predominantly white
Interventions	Total number of intervention groups: 2 Control group description: no intervention Total number of participants randomised to each of the groups: 34 to control, 33 to RA‐led distraction, 32 to parent‐led distraction Specific intervention #1: RA‐led distraction Intervention details: research assistant held toy and distracted infant Integrity of intervention: good Specific intervention #2: parent‐led distraction Intervention details: parent distracted infant throughout procedure Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: MBPS Time points: 15 seconds to 1 minute after needle Upper limit: (variable): 10 Lower limit: 0 seconds Higher score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of the study authors: toy distraction does not work Study funding source: this project was funded by a SSHRC to Ms Jessica Hillgrove‐Stuart and a CIHR New Investigator Award to Dr R Pillai Riddel Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	High risk	Personnel and participants were not blinded.
3b) Detection bias (blinding of outcome assessors)	High risk	Coders were not blinded.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (2/7), unclear risk (1/7).

Ho 2016.

Study characteristics
Methods	Study design: between‐group Total study duration: 3 months
Participants	Total number: 54 Setting: hospital (NICU) Diagnostic criteria: stable preterm neonates Gestational age: Mean: 34.2 (swaddling), 34.5 (control) SD: 1.5 (swaddling), 1.9 (control) Postnatal age: Mean: 9.3 days(swaddling), 11.2 days (control) SD: 9.1 (swaddling), 14 (control) Sex: Males: 31 Females: 23 Country: Hong Kong
Interventions	Total number of intervention groups: 1 Control group description: standard care in supine position Total number of participants randomised to each of the groups: Intervention: 27 Control: 27 Intervention descriptions: the infants were placed in a supine position on a blanket while their arms were flexed, crossed‐over, and close to the chest. The end of one side of the blanket was folded over the neonate’s chest and arms to the opposite side, and tucked under the trunk. This folding procedure continued for the end of the other side of the blanket and tucked under the opposite side of the trunk. To ensure that the blanket was not over restrictive, the blanket was checked by sliding a finger between the wrap and the neonate. The lower portion of the blanket was folded up to the belly with one of neonate’s feet uncovered. This was done to expose the neonate’s heel for heel stick procedure. Then, the neonate was placed in a lateral position. The blanket should be able hold the neonate firmly but gently in a flexed position. The swaddling procedure took about 1 minute. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure and 2 minutes post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: the findings show that swaddling is feasible and efficacious in controlling pain for heel stick procedure among preterm infants Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Authors address the inability to blind study personnel to study conditions given the nature of the intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	There is no mention of outcome assessor being blinded. Additionally, given that assessment happened in the moment, it would not have been possible to blind the assessor due to the nature of the intervention.
4) Attrition bias (incomplete outcome data)	Low risk	Authors note that 27 infants were recruited to each group and that no participants dropped out or were lost to follow up.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	High risk	All the infants in the swaddled group were placed in a lateral position while all the infants in the control were in supine position. Authors mention this as a limitation since it is hard to tease apart the effects of swaddling and positioning.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall assessment is high risk.

Hogan 2014.

Study characteristics
Methods	Study design: between‐group Total study duration: 15 months
Participants	Total number: 120 Setting: primary care paediatric practice Diagnostic criteria: Inclusion criteria: aged 4 to 6 months scheduled to have their DTaP‐IPV‐Hib and PCV vaccines Exclusion criteria: if impaired neurological development, history of seizures, use of local topical analgesics at injection site, use of sedatives or opioids 24 hours prior to procedure, fever or illness that would prevent administration of vaccination, a parent who was unable to use the assessment tool in the study, or a parent who did not speak English. Each infant could only participate once. Age: Minimum: 4 months Maximum: 6 months Mean % of each age group: 4 months = 71%; 6 months = 29% SD: unknown Sex: Males: 61 Females: 59 Country: Canada
Interventions	Total number of intervention groups: 1 Control group description: no tactile stimulation. Note: all infants received sucrose solution, holding by parent, and injections without aspiration. Total number of participants randomised to each of the groups: Intervention: 60 Control: 60 Specific intervention: parent‐led tactile stimulation Intervention details: rub ipsilateral thigh distal to injection site for 15 seconds prior, during, and after each needle, using as much pressure as they felt would be suitable without distressing their infant Integrity of intervention: good
Outcomes	Pain outcome, scale limits and time points when measured: MBPS Time points: 15‐second intervals before and after each injection Upper limit: 10 Lower limit: 0 High score = more pain Continuous VAS parent rating Time points: during each injection Upper limit: unknown Lower limit: unknown High score = more pain Duration of crying Time points: first minute after each vaccine VAS RA rating Time points: during each injection Upper limit: unknown Lower limit: unknown High score = more pain Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: parent‐led tactile stimulation did not reduce pain in infants undergoing immunisation when combined with other pain relieving interventions (sucrose and parenting holding) Study funding source: M‐E.H. was supported during this study by a Pain in Child Health Training Consortium award and a CIHR Frederick Banting and Charles Best Canada Graduate Scholarship‐ Master's Award. R.P.R. holds a CIHR New Investigator Award and an Ontario Ministry of Research and Innovation Early Researcher Award. A.T. and R.P.R. hold a research grant from Pfizer Canada Inc. (Toronto, Canada) Author COI statement: the remaining authors declare no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	“A computer random number generator was used to create the sequence in random blocks of 4 to 8” (page 2).
2) Selection bias (allocation concealment)	Low risk	“The randomization code was created by an individual not directly associated with the study. Treatment allocation was concealed in sequentially numbered opaque sealed envelopes that were not opened until after written consent was obtained from a parent” (page 2).
3a) Performance bias (blinding of personnel)	Unclear risk	Participants were blinded but personnel were not.
3b) Detection bias (blinding of outcome assessors)	High risk	Coders were blinded to study hypotheses but not blinded to condition.
4) Attrition bias (incomplete outcome data)	Low risk	A reasonable account of how attrition was dealt with is given.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (1/7), unclear risk (1/7).

Im 2008.

Study characteristics
Methods	Study design: randomised between groups Total study duration: not reported
Participants	Total number: 99 Setting: hospital Diagnostic criteria: healthy neonates Age: Minimum: 38.8 weeks Maximum: 39.1 weeks Mean: approximately 273.4 days gestation SD: approximately 9.3 days gestation Sex: Males: 47 Females: 52 Country: South Korea Socio‐demographics: unknown Ethnicity: Korean
Interventions	Total number of intervention groups: 2 Control group description: one Total number of participants randomised to each of the groups: touch: 33; non‐nutritive sucking: 33; control: 33 Specific intervention #1: Yakson touch Intervention details: nurse warmed hands to 34 degrees Celsius and approached neonate from behind. Nurse placed left hand under neonate's back. Nurse placed right hand on neonate's abdomen and caressed clockwise (approximately 4 cm in diameter) every 5 seconds. Integrity of intervention: good Specific intervention #2: non‐nutritive sucking Intervention details: neonates were given a small, short, and hollow nipple packed with sterile gauze while gentle pressure was applied in the mouth Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: NIPS Time points: 1 minute post heel stick Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: touch and non‐nutritive sucking were found not to reduce pain in healthy neonates measured 1 minute after heel stick, as compared to the control group Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Unclear risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (3/7), high risk (1/7), unclear risk (3/7).

Ipp 2004.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 106 Setting: clinic Diagnostic criteria: healthy infants Age: M = 123 days, SD = 51 Sex: Males: 47 Females: 59 Country: Canada
Interventions	Total number of intervention groups: 1 (held by parent) Control group details: infants were placed unrestricted on a clean paper towel on the examination table. The mothers were instructed to interact with their infants in the way they were used to and which felt most comfortable and familiar to them. Total number of participants randomised to each of the groups: held by parent (HD) = 56, supine (SUP) = 50 Specific intervention: held by parent (HD), supine (SUP) Intervention details: mothers were instructed to stand and hold the baby in a comfortable way (no specific instructions given) Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured NFCS Time points: not reported Upper limit: 300% Lower limit: 0% High score = more pain Continuous Crying 180 seconds Upper limit: 180 seconds Lower limit: 0 More crying = more pain Continuous VAS 15 seconds post needle Upper limit: 100 mm (maximum pain) Lower limit: 0 mm (no pain) High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: holding is not better than the supine positioning in reducing the post needle pain in healthy infants aged 2 to 6 months (based on crying and NFCS and VAS scores) Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported but this did not involve an independent person.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	High risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score on Cochrane standard: low risk (4/7), high risk (1/7), unclear risk (2/7).

Jain 2006.

Study characteristics
Methods	Study design: cross‐over Total study duration: not reported
Participants	Total number: 23 Setting: hospital (NICU) Diagnostic criteria: hospitalised preterm newborns Gestational age: Mean: 31.1 weeks SD: 1.9 Minimum: 28 weeks Maximum: 35 weeks Sex: not reported Country: Canada
Interventions	Total number of intervention groups: 1 Description of control group: control group did not receive any pain management strategies   Total number of participants randomised to each of the groups: 23 Specific intervention: massage Intervention details: the infant was wrapped and bundled for 5 minutes prior to heel lance. The heel (from toes to mid thigh) was massaged with gentle pressure using fingers and thumbs. The heel was then covered with a warm cloth for 2 to 3 minutes. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: 5 minutes post heel lance Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of the study authors: gentle massage of the leg prior to heel stick is safe and decreases pain responses in preterm infants Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Unclear risk	Efforts to reduce measurement bias are noted but not explained.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	It is not clear whether the outcome assessor was blinded to the intervention.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, if one or more domain is "unclear", the overall judgement is unclear.

Jebreili 2015.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 135 Setting: hospital (NICU) Diagnostic criteria: clinically stable preterm infants Gestational age: Mean: 31.6 (breast milk odour), 30.9 (vanilla odour), 31.5 (control), 31.3 (overall) SD: 2.1 (breast milk odour), 2.0 (vanilla odour), 2.0 (control) Postnatal age: Mean: 10.0 days (breast milk odour), 11.4 days (vanilla odour), 11.0 days (control), 10.7 (overall) SD: 7.1 (breast milk odour), 7.1 (vanilla odour), 7.9 (control) Sex: Males: 62 Females: 73 Country: Iran
Interventions	Total number of intervention groups: 2 Description of control group: not reported Total number of participants randomised to each of the groups: breast milk odour = 45; vanilla odour = 45; control = 45 Specific intervention #1: breast milk odour Intervention details: infants in the breast milk group were exposed to breast milk odour (their own mother’s milk) 5 minutes before the start of sampling to 30 seconds after the end of sampling. During the procedure all infants were kept at 37 degrees C. Integrity of intervention #1: poor Specific intervention #2: vanilla odour Intervention details #2: infants in the vanilla group were exposed to familiar vanilla odour 5 minutes before the start of sampling to 30 seconds after the end of sampling. Infants of the vanilla group were familiarised with vanilla odour by a 10 g clean cotton ball dipped in 10 drops of vanilla solution. Cotton balls were placed in the incubators 10 cm from the infant’s nose 12 hours before sampling. Integrity of intervention #2: poor
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure and 1 minute post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of the study authors: breast milk odour can be used for calming premature infants during and after venipuncture Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	Unclear risk	It is mentioned that the intervention happened during and after venipuncture, but not enough details are provided about who is conducting the procedure and whether they are blinded to study condition.
3b) Detection bias (blinding of outcome assessors)	Low risk	PIPP scores coded from videos so outcome assessors could not have smelt the 2 different smells.
4) Attrition bias (incomplete outcome data)	Low risk	45 infants were recruited for each group (135 total) and data on 45 infants for each group were analysed and reported.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, if one or more domain is "unclear", the overall judgement is unclear.

Johnston 1997.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 44 Setting: NICU Diagnostic criteria: premature sample Age: Minimum: 25 4/7 weeks Maximum: 34 2/7 weeks Mean: 31 weeks SD: not reported Sex: Males: 42 Females: 43 Country: Canada Co‐morbidity: no congenital abnormalities; Apgar scores greater than 3 to 5 minutes Socio‐demographics: not reported Ethnicity: not reported
Interventions	Total number of intervention groups: 1 Control group description: baby was handled and repositioned to the side, and 0.05 mL of water was placed on the tongue's surface just prior to starting the heel lance procedure Total number of participants randomised to each group: 24 in rocking condition 20 in control condition Specific intervention: rocking Intervention details: the infant was repositioned to be on the side or in the supine position and was swaddled in a blanket. An air mattress was placed beneath them, that pumped air into the mattress at a rate of 12 cycles per minute. The swaddled baby was left on the oscillating mattress for 15 minutes prior to the heel stick. The infant also received 0.05 mL of water. Integrity of intervention: good
Outcomes	Pain outcomes, time points when measured, and scale limits: NFCS Baseline, 30 seconds post needle, 60 seconds post needle, 90 seconds post needle Upper limit: 1 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions: no difference between rocking and control Study funding source: funding for this project was primarily by the Fonds de Recherche en Sante au Quebec (FRSQ940989‐104) with additional funding from National Institute of Nursing Research of the NIH (NR03916 ROI) and for salary support from Health Canada (C.J.) and the Ontario Ministry of Health (B.S.) Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Unclear risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	High risk	None specified and suspect coders were not blinded due to rocking in an incubator.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (1/7), unclear risk (2/7).

Johnston 2007a.

Study characteristics
Methods	Study design: cross‐over Total study duration: not reported
Participants	Total number: 19 Setting: NICU Diagnostic criteria: preterm infants Age: Minimum: 32 weeks Maximum: 34 5/7 weeks Mean: 33 weeks Std. 0.9 weeks Sex: Males: 57% males Females:43% females Country: Canada
Interventions	Total number of intervention groups: 1 Control group description: no intervention Total number of participants randomised to each of the groups: 19 Specific intervention: mother's voice played during heel lance Intervention details: recording of mother's voice either singing, talking, or reading a nursery rhyme was played 1 minute before the procedure began and until heart rate and O2 sat returned to baseline Integrity of intervention: good
Outcomes	Pain outcomes, time points when measured, and scale limits: PIPP 30 seconds Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of the authors: playing maternal voice pre and during heel lance has no significant impact on preterm pain Miscellaneous comments by review authors: no means, only F provided for PIPP Percentages for NFCS Study funding source: funded by Canadian Institutes of Health Research MOP‐38074 and in part by McGill University James McGill Chair Fund Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Unclear risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	Low risk	Coders blinded to the purpose of the study.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (1/7), unclear risk (2/7).

Johnston 2013.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 55 Setting: hospital Diagnostic criteria: preterm neonates between 26 0/7 and 28 6/7 weeks’ gestational age. Infants were free of major congenital anomalies, had not experienced severe hypoxia, and were not receiving analgesics or sedatives within 72 hours of the study; however, they could be receiving ventilator support Gestational age: M = 27 2/7 weeks   Sex: Intervention: 53% male, 47% female Control: 47% male, 53% female Country: Canada
Interventions	Total number of intervention groups: 1 Control group description: sham group: the therapist stood beside the incubator with hands at her side and did mental tasks (e.g. mathematical operations) to avoid focusing on the infant Total number of participants randomised to each of the groups: Therapeutic touch: 27 Control: 28 Specific intervention: therapeutic touch Intervention details: the therapist used her hands to assess and re‐balance the energy field of the patient using the following steps: (1) centring her state of awareness; (2) assessing the energy field of the patient; and (3) modulating the energy field Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: during heel lance ‐ 30, 60, 90, 120 seconds Upper limit: 21 Lower limit: 3 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: therapeutic touch given immediately before and after heel lance has no comforting effect in preterm neonates. Other effective strategies involving actual touch should be considered. Study funding source: funded by the SickKids Foundation, Toronto (CAM 06‐315), and infrastructure support for the laboratory was provided by Canadian Institutes of Health Care, the Quebec Health Research Fund, Ottawa, and the Quebec Interuniversity Nursing Research Group, Montreal, Canada Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Unclear whether what was used was a website for random numbers.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	A convincing effort to reduce bias in outcome measurement is reported.
3b) Detection bias (blinding of outcome assessors)	Low risk	Coders were blinded.
4) Attrition bias (incomplete outcome data)	Low risk	For outcome examined there were sufficient outcome data.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score on Cochrane standard: low risk (5/7),High risk (0/7), unclear risk (2/7).

Jose 2012.

Study characteristics
Methods	Study design: quasi‐experimental post‐test only design Total study duration: not reported
Participants	Total number: 60 Setting: various rural maternal and child welfare centres of Kasturba Medical College, Manipal, India Diagnostic criteria: not reported Age: all infants were 14 weeks old Minimum: 14 weeks Maximum: 14 weeks Mean: 14 weeks SD: 0 weeks Sex: Males: 33 Females: 27 Country: India
Interventions	Total number of intervention groups: 1 Control group description: no intervention Total number of participants randomised to each of the groups: Intervention: 30 Control: 30 Specific intervention: skin‐tap technique Intervention details: tapping the skin with the finger pads of the dominant hand to relax the muscle. The immunisation site (i.e. left vastus lateralis) was tapped for a duration of 2 minutes before immunisation, and during and after immunisation, the area above the site was tapped for a duration of 1 minute. To maintain synchrony, the investigator counts 1, 2 and on the count of 3, the needle will be inserted. Integrity of intervention: satisfactory
Outcomes	Pain outcome, scale limits and time points when measured: Behavioral Observation Pain Scale Time points: not specified Upper limit: 20 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: the study concludes that skin tapping was effective in improving pain during DPT injections among infants. The study recommends application of this technique to nursing practice to minimise the pain experienced by the children they are taking care of.
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	High risk	"Quasi experimental post‐test only control group", "the sampling design was purposive sampling with random allocation of treatment using chit method with non replacement technique".
2) Selection bias (allocation concealment)	Unclear risk	Not specified.
3a) Performance bias (blinding of personnel)	Unclear risk	Not specified.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	Not specified.
4) Attrition bias (incomplete outcome data)	Unclear risk	It was unclear how missing participants or data were accounted for.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Unclear risk	Not specified.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (0/7), high risk (1/7), unclear risk (6/7).

Kahraman 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: 20 months
Participants	Total number: 64 Setting: hospital Diagnostic criteria: premature hospitalised infants, but described as stable Gestational age (weeks): Mean: 33.8 (white noise), 34.0 (mother's recorded voice), 34.1 (MiniMuffs), 34.3 (control), overall mean = 34.0 SD: 1.8 (white noise), 1.5 (mother's recorded voice), 1.8 (MiniMuffs), 1.7 (control), overall SD = 1.6   Sex: Males: 7 (white noise), 8 (mother's recorded voice), 8 (MiniMuffs), 9 (control), total = 32 Females: 9 (white noise), 8 (mother's recorded voice), 8 (MiniMuffs), 7 (control), total = 32 Country: Turkey
Interventions	Total number of intervention groups: 3 Control group description: administered standard care Total number of participants randomised to each of the groups: white noise = 16, mother's recorded voice = 16, MiniMuffs = 16, control = 16 Specific intervention: white noise, mother's recorded voice, MiniMuffs Intervention details: white noise: the white noise was started 5 minutes before the heel lance and was played to the infant during the procedure. Dr. Harvery Karp’s "The Happiest Baby", which consists of only intrauterine sounds, was used. The speaker was placed on the tip of the foot at a distance of about 30 cm from the ear of the neonate 5 minutes before the procedure and then played to the infant during the procedure. The sound level was adjusted to 50 dB using the Benetech Digital Sound Level Meter for the mother’s voice and white noise groups. Mother's recorded voice: The mothers of the infants were given voice recorders and asked to record their voice in a comfortable room saying whatever they wanted to their infant. Each mother recorded her voice for 5 minutes. The voice recorder was placed on the tip of the foot at a distance of about 30 cm from the ear of the neonate 5 minutes before the procedure and then played to the infant during the procedure. The sound level was adjusted to 50 dB using the Benetech Digital Sound Level Meter for the mother’s voice and white noise groups. MiniMuffs: The infants had MiniMuffs placed on their ears 5 minutes before the procedure to reduce the environmental noise. Natus MiniMuffs, Neonatal Noise Attenuators have been developed for newborns and premature infants. MiniMuffs protect the sensitive ears of the premature and provide a safe environment for healthy development. MiniMuffs reduce sound levels by at least 7 dB and allow the sound pressure level to decrease by more than 50%. Integrity of intervention: white noise = good, mother's recorded voice = good, MiniMuffs = good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: during heel lance Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: auditory interventions used during heel lance reduce the pain and increase the comfort of the premature infants. White noise is extremely effective in preventing infants’s pain. Study funding source: this work was supported by Ege University Scientific ResearchProjects Directorate Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Low risk	A convincing method for allocation concealment is reported.
3a) Performance bias (blinding of personnel)	High risk	Blinding not possible due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors were stated to be unblinded to group allocation.
4) Attrition bias (incomplete outcome data)	Low risk	Showed attrition reasons, which appear balanced across groups.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

KarabiyikOgurlu 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 80 Setting: hospital Diagnostic criteria: healthy full‐term newborns Age (days): Median: one (warm compression), one (control), overall = 1 Range: one to 2 (warm compression), one to 2 (control), overall = 1 to 2 Sex: Males: 22 (warm compression), 20 (control), total = 42 Females: 18 (warm compression), 20 (control), total = 38 Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: no comforting or relaxation interventions were used during the heel lance procedure, and these newborns were immediately comforted after the procedure Total number of participants randomised to each of the groups: Intervention: 40 Control: 40 Specific intervention: warm compression Intervention details: before the heel lance procedure, the newborns were subjected to a local dry mildly warm compress for 5 minutes using a thermophore. The warmth of the water in the thermophore was kept between 34 ºC and 37 ºC. To prevent the thermophore from contacting the sole of the infant’s foot, it was wrapped in a cloth and placed on the sole from which heel lance would be taken. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: during the procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: it was found that warm compression applied before heel lance decreases the length of the procedure and is effective in decreasing pain and providing comfort in newborns Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel (i.e. nurse) were not blinded to group allocation.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	It is unclear if outcomes assessors were blinded.
4) Attrition bias (incomplete outcome data)	Unclear risk	Attrition/exclusions not described so unable to determine whether there were incomplete outcome data. It was unclear how missing participants or data were accounted for.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Karaca 2016.

Study characteristics
Methods	Study design: between‐group Total study duration: 4 months
Participants	Total number: 70 Setting: family health centre Diagnostic criteria: healthy full‐term infants Gestational age: Mean: 39.9 (flick), 39.7 (control), 39.8 (overall) SD: 0.6 (flick), 0.7 (control), 0.7 (overall) Sex: Males: 37 Females: 33 Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: the infants in the control group were regularly vaccinated, without any other attempt in FHC, and they were videotaped before, during, and after the vaccination. During the vaccination procedure, the parents of infants in both groups were allowed to calm their babies by touching and talking to them, but not to feed and do anything that would distract the infant’s attention, including giving them toys, showing them a dummy, or clapping. Total number of participants randomised to each of the groups: Flick: 35 Control: 35 Specific intervention: flick Intervention details: the nurse determined the vaccination area and disinfected it using cotton with 70 % alcohol. The muscle was held with the nurse’s left hand, and the vaccination area was flicked with the right hand. The flick was given as follows: the thumb was placed on the nail of middle finger, then the vaccination area was stimulated with a quick tap using upper nail part of middle finger. It was ensured that the hands of flicking staff were warm, and nails were clipped in order not to harm the infant. The staff waited to start the procedure until the infant calm down, if it was anxious; the procedure was not initiated if the infants were crying. During the vaccination procedure, the parents of infants in both groups were allowed to calm their babies by touching and talking to them, but not to feed and do anything that would distract the infant’s attention, including giving them toys, showing them a dummy, or clapping. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time point: first 30 seconds post‐procedure and 1 minute post‐procedure High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of the study authors: flick application at vaccination area could be used to reduce pain during vaccination at babies Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	High risk	Cochrane states that "Sequence generated by some rule based on date (or day) of admission" is considered high risk, which is analogous to the method described in the article.
2) Selection bias (allocation concealment)	High risk	Description analogous to Cochrane's example of "Alternation or rotation" so high risk of bias.
3a) Performance bias (blinding of personnel)	High risk	Unable to blind nurse that is administering vaccination given that they are the one implementing the intervention (flick).
3b) Detection bias (blinding of outcome assessors)	Unclear risk	Although it's stated that outcome assessors were blinded, to code the NIPS from a video you need to get a full body video and so it is unclear whether the flick on the upper thigh would have been caught on camera.
4) Attrition bias (incomplete outcome data)	Low risk	70 infants were randomised and 70 infants were included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall assessment is high risk.

Khasanah 2019.

Study characteristics
Methods	Study design: between‐group  Total study duration: not mentioned
Participants	Total number: 47 Setting: community health centre Diagnostic criteria: healthy neonates   Postnatal age (months):  2 to 12 Mean: not mentioned  SD: not mentioned  Median: 4 (intervention by mother), 4 (intervention by health worker), 3 (control)   Sex:  Males: not mentioned Females: not mentioned Country: Indonesia
Interventions	Total number of intervention groups: 2 Control group description: participants received sitting up position by mother when intramuscular injection   Total number of participants randomised to each of the groups: Mother: 16 Health worker: 16 Control: 16   Specific intervention #1: mother  Intervention details #1: participants received facilitated tucking combined with sitting up position by mother when intramuscular injection. Integrity of intervention #1: poor   Specific intervention #2: Health worker  Intervention details #2: participants received facilitated tucking combined with sitting up position by health worker when intramuscular injection Integrity of intervention #2: poor
Outcomes	Pain outcomes, scale limits, and time points when measured FLACC Time points: during procedure Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors: intervention by mother group had the lowest level of pain compared to the others with the level found to be moderate with intervention by health workers. No significant differences in pain levels were found between the control and intervention by mother groups. It is recommended that facilitated tucking by mother combined with sitting up position should be used in infant pain management during intramuscular immunisation.  Study funding source: not reported Author COI statement: there are no conflicts of interest in this study
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Authors do not sufficiently describe the method used to generate the randomisation.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	Unclear risk	It is not clear whether the personnel were blinded to the intervention.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were blinded to study condition.
4) Attrition bias (incomplete outcome data)	Low risk	The number of enrolled participants in each group and the number reported in outcome tables match.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, if one or more domain is "unclear", the overall judgement is unclear.

Khosravan 2018.

Study characteristics
Methods	Study design: between‐group Total study duration: 6 months
Participants	Total number: 90 Setting: clinic Diagnostic criteria: healthy full‐term infants (2 months) Postnatal age: Mean: 60.2 days (massage with ice), 61.0 days (massage without ice), 61.3 days (control), 60.8 days (overall) Sex: Males: 45 Females: 45 Country: Iran
Interventions	Total number of intervention groups: 2 Description of control group: infants in the control group put on the bed in supine position without any intervention and based on the usual procedure of the centre   Total number of participants randomised to each of the groups: massage with ice = 30; massage without ice = 30; control = 30 Specific intervention #1: massage with ice Intervention details: Hugo point of the right hand massaged for 1 minute using ice cube covered in plastic glove and 1‐layer cotton cloth; intervention is given for 1 minute, followed by 30 seconds of relaxation, and then another 1 minute of intervention administration Integrity of intervention #1: good Specific intervention #2: massage without ice Intervention details #2: Hugo point of the right hand massaged for 1 minute using soft part of the thumb; average amount of pressure applied; intervention is given for 1 minute, followed by 30 seconds of relaxation, and then another 1 minute of intervention administration Integrity of intervention #2: good
Outcomes	Pain outcomes, scale limits, and time points when measured MBPS Time point: first 30 seconds post‐procedure Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of the study authors: Hugo’s point massage with and without ice can reduce the injection‐related pain of pentavalent vaccine in infants and be used as a simple non‐pharmaceutical method available to control vaccination pain Study funding source: Gonabad University of Medical Sciences
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel unable to be blinded due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors unable to be blinded due to nature of intervention.
4) Attrition bias (incomplete outcome data)	Low risk	The number of enrolled participants in each group and the number reported in outcome tables match.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain is "high", the overall judgement is high.

Kozub 2001.

Study characteristics
Methods	Study design: cross‐over Total study duration: winter of 2000
Participants	Total number: 20 Setting: pulmonary clinic (hospital) Diagnostic criteria: premature born infants (but aged 1 to 15 months at time of the study), receiving Synagis injections to prevent pulmonary infections Age: Minimum: 1 month Maximum: 15 months Mean: 6.6 months SD: 4.1 months Sex: Males: 12 Females: 8 Country: USA
Interventions	Total intervention groups: 1 Control group description: 1 (mimic intervention) Total number of participants randomised to each of your groups: 20 Specific intervention: therapeutic touch Intervention details: Krieger method: centring, assessment, unruffing and direction/modulation of energy Centring: focus attention Assessment: detection of irregular energy areas (chakras) Unruffing: smoothing at energy field Direction/modulation of energy: energy transfer (change flow) and using visualisation and intentionality to redistribute energy flow Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP 30 seconds post injection Upper limit: 21 Lower limit: 0 High score = more distress Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Funding sources: not reported Key conclusions of study authors: therapeutic touch does not reduce immediate (30 seconds) pain during heel stick in infants aged 1 to 15 months
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	A convincing effort to reduce bias in outcomes measurement is reported.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score on Cochrane standard: low risk (5/7), high risk (0/7), unclear risk (2/7).

Koç 2015.

Study characteristics
Methods	Study design: between‐group Total study duration: 5 months
Participants	Total number: 60 Setting: family health centre Diagnostic criteria: healthy full‐term infants Gestational age: Mean: 39.1 (reflexology), 39.3 (control), 39.2 (overall) SD: 0.9 (reflexology), 0.8 (control), 0.8 (overall) Postnatal age: Mean: 126.3 days (reflexology), 125.9 days (control), 126.1 days (overall) SD: 95.6 (reflexology), 97.1 (control), 95.5 (overall) Sex: Males: 28 Females: 32 Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: no description Total number of participants randomised to each of the groups: Reflexology: 30 Control: 30 Specific intervention: reflexology Intervention details: before reflexology, the practitioner’s hands were washed and kept at body temperature. The treatment was conducted while the infant was in his or her mother’s arms to make the infant relax and feel more secure. Reflexology was performed starting on the toes of one foot down to the heel; the same process was performed on the other foot. Reflexology was performed for an average period of 20 to 30 minutes according to the physical size of the infant’s feet. The researcher performing reflexology treatment had successfully completed 100 hours of theoretical and applied training before starting the study and received their certification before starting the study. Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured FLACC Pain Scale Time point: first 30 seconds post‐procedure Upper limit: 10 Lower limit: 0 High score = higher pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of the study authors: reflexology before vaccine reduced the pain level experienced after vaccination Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	Envelopes were sealed, but no mention of whether they were sequentially numbered or opaque. There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	Unclear risk	It is not clear whether the person administering the vaccine was blinded to study condition. The family health centre nurse administered the vaccines (and did not administer the intervention) and could have entered the room after reflexology was administered, but that is not specified. More information is required.
3b) Detection bias (blinding of outcome assessors)	High risk	Need video of whole body to code FLACC so impossible to blind outcome assessor to study condition (i.e. in reflexology group).
4) Attrition bias (incomplete outcome data)	Unclear risk	2 dropouts because of reasons related to intervention. Since Ns were equal after these 2 participants dropped out, unclear if this had an impact. It was unclear how missing participants or data were accounted for.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall assessment is high risk.

Kristoffersen 2011.

Study characteristics
Methods	Study design: cross‐over Total study duration: 42 months; each study participant was examined over a 3‐week study period
Participants	Total number: 24 Setting: the NICU at St. Olav’s University Hospital Diagnostic criteria: infants with a gestational age at time for first intervention (i.e. postmenstrual age or PMA) between 28 and 32 weeks, in stable condition, and at low risk for neurologic sequelae. Infants on a ventilator or those receiving continuous positive airway pressure (CPAP) or opioid treatment were excluded, as were infants with ongoing serious infections. Infants who needed to have their tube inserted by mouth were not included. Age: Minimum: PMA of 28 weeks at first intervention Maximum: PMA of 32 weeks at first intervention Mean: not reported SD: not reported Sex: Males: not reported Females: not reported Country: Norway
Interventions	Total number of intervention groups: 5 Control group description: no intervention Total number of participants randomised to each of the groups: intervention #1 = 24; intervention #2 = 24; intervention # 3 = 24; intervention #4 = 24; intervention #5 = 24; control = 24 Specific intervention #1: no pacifier + sterile water Specific intervention #2: no pacifier + sucrose 30% Specific intervention #3: pacifier + no fluid Specific intervention #4: pacifier + sterile water Specific intervention #5: pacifier + sucrose 30% Integrity of intervention: good
Outcomes	Pain outcome, scale limits, and time points when measured: PIPP Time points: during nasogastric tube insertion. After 1 minute, and after 5 minutes. Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: the intervention that best reduced pain reactivity was a pacifier with sucrose Study funding source: the study was supported by grants from Unimed Innovation, St Olav’s University Hospital, Trondheim, Norway Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	High risk	Personnel were not blinded.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Unclear risk	It was unclear how missing participants or data were accounted for.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (1/7), unclear risk (2/7).

Kucukoglu 2015.

Study characteristics
Methods	Study design: between‐group Total study duration: 4 months
Participants	Total number: 60 Setting: hospital (NICU) Diagnostic criteria: full‐term Gestational age: Intervention: 38.7 ± 0.8 weeks Control: 38.9 ± 0.8 weeks   Sex: Intervention: 19 females, 11 males Control: 15 females, 15 males Overall: 34 females, 26 males Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: no application was performed on the newborns in the control group. Vaccination was performed in the supine position on infants in this group as routine clinical practice. The leg that received the vaccination was brought to a straight position (classical holding position), and 70% alcohol was used to clean the area to be vaccinated in accordance with the clinical protocol. Total number of participants randomised to each of the groups: Intervention: 30 Control: 30 Specific intervention: facilitated tucking Intervention details: each newborn in the treatment group was prepared in the facilitated tucking position 1 minute before the procedure by the assistant nurse, and 70% alcohol was used to clean the area to be vaccinated in accordance with the clinical protocol. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: during procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: the results showed that the facilitated tucking position is an effective method for reducing pain during the procedure Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	Authors describe a randomisation method similar to "Shuffling cards or envelopes".
2) Selection bias (allocation concealment)	Unclear risk	While authors mention sealed envelopes, there is not enough information to assess allocation concealment. There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	No mention of blinding and also unable to blind study personnel due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors for the NIPS were not aware of study condition.
4) Attrition bias (incomplete outcome data)	Low risk	60 infants were enrolled and randomised to study conditions, and 60 infants were included in analyses. Therefore, there are no missing outcome data.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall assessment is high risk.

Leng 2016.

Study characteristics
Methods	Study design: between‐group Total study duration: 8 months
Participants	Total number: 671 Setting: hospital (NICU) Diagnostic criteria: healthy full‐term neonates (between 37 to 40 weeks gestational age) Gestational age: Median: 37.6 (sucrose + warmer), 37.9 (sucrose + NNS warmer), 37.8 (sucrose + swaddling), 37.7 (sucrose + NNS + swaddling) Postnatal age: Median: 8.1 (sucrose + warmer), 8.1 (sucrose + NNS + warmer), 8.5 (sucrose + swaddling), 7.6 (sucrose + NNS + swaddling) Range: 3 to 20 days (sucrose + warmer), 3 to 23 days (sucrose + NNS+ warmer), 3 to 26 days (sucrose + swaddling), 3 to 21 days (sucrose + NNS + swaddling) Sex: Males: 50.8% (sucrose + warmer), 52.7% (sucrose + NNS + warmer), 58.4% (sucrose + swaddling), 50.9% (sucrose + NNS + swaddling) Females: 49.2% (sucrose + warmer), 47.3% (sucrose + NNS + warmer), 41.6% (sucrose + swaddling), 49.1% (sucrose + NNS + swaddling) Country: Iran
Interventions	Total number of intervention groups: 4 Total number of participants randomised to each of the groups: sucrose + warmer = 176, sucrose + NNS + warmer = 167, sucrose + swaddling = 167, sucrose + NNS + swaddling = 161 Specific intervention #1: sucrose + warmer Intervention details: 2 mL of 24% sucrose was administrated to the infant’s mouth by syringe 2 minutes before the heel stick procedure. Procedures were performed 1 hour after feeding. Neonates who were wearing baby clothing were placed under the radiant warmers before heel stick procedure with the temperature set to 36°C. Integrity of intervention #1: good Specific intervention #2: sucrose + NNS + warmer Intervention details #2: 2 mL of 24% sucrose was administrated to the infant’s mouth by syringe 2 minutes before the heel stick procedure, and then a standard silicone newborn pacifier was placed into the infant’s mouth until the end of the process. Procedures were performed 1 hour after feeding. Neonates who were wearing baby clothing were placed under the radiant warmers before heel stick procedure with the temperature set to 36°C. Integrity of intervention #2: good Specific intervention #3: sucrose + swaddling Intervention details #3: infants were swaddled with a cotton blanket, upper but not lower limb movements were restricted by the blanket, and then 2 mL of 24% sucrose was administrated to the infant’s mouth by syringe 2 minutes before the heel stick procedure. The lower limbs were swaddled right after the heel stick procedure until the end of the process. Neonates were not exposed to warm temperatures during this procedure. Integrity of intervention #3: good Specific intervention #4: sucrose + NNS + swaddling Intervention details #4: infants were swaddled with a cotton blanket, upper but not lower limb movements were restricted by the blanket, then 2 mL of 24% sucrose was administrated into the infant’s mouth by syringe before the heel stick procedure, then a standard silicone newborn pacifier was placed into the infant’s mouth, the lower limbs were swaddled right after the heel stick procedure until the end of the process. Neonates were not exposed to warm temperatures during this procedure. Integrity of intervention #4: good
Outcomes	Pain outcomes, scale limits, and time points when measured: NFCS‐Revised Time point: 1 minute post‐procedure Upper limit: 5 Lower limit: 0 High score = higher expression of pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of the study authors: non‐nutritive sucking and swaddling had synergistic effects on pain relief when used with oral sucrose. For the deep heel stick procedure, oral sucrose combined with NNS and swaddling provided the best pain relief effect. For the shallow heel stick procedure, addition of NNS and swaddling did not improve the effects. Study funding source: national key clinical specialist construction Program of China and Medical Research Funding of Health Bureau of Chongqing Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes potential biases of investigators.
3a) Performance bias (blinding of personnel)	High risk	While the authors state that the nurses performing heel sticks and sucrose administration were blinded to the study and infants' clinical information, they could not have been blinded to whether or not the baby was swaddled.
3b) Detection bias (blinding of outcome assessors)	High risk	No mention of blinding of outcome assessors, and blinding very difficult to ensure given the nature of the intervention (i.e. swaddling).
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall rating is high risk.

Liaw 2010.

Study characteristics
Methods	Study design: between‐group Total study duration: 9 months
Participants	Total number: 104 Setting: NICU in Taiwan Diagnostic criteria: healthy, preterm infants, between 3 and 28 days post birth Age: 27 to 37 weeks Minimum: 3 days Maximum: 19 days Mean: 6.5 days SD: 3.1 days Sex: Males: 54 Females: 50 Country: Taiwan
Interventions	Total intervention groups: 1 Control group description: routine comfort (gentle touching) without non‐nutritive sucking Total number of participants randomised to each of the groups: 52 Specific intervention: non‐nutritive sucking Intervention details: pacifier was given to infant to suck before touching the foot to initiate the heel stick Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured: PIPP Every minute for before (for 3 minutes), during, and after (for 10 minutes) the procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: non‐nutritive sucking reduces pain, especially mild to moderate pain and behavioural responses Study funding source: this research was supported by a grant from Tri‐Service General Hospital, Taipei, Taiwan Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	A convincing effort to reduce bias in outcomes measurement is reported.
3b) Detection bias (blinding of outcome assessors)	High risk	None specified but suspect that coders were not blinded due to sucking.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (1/7), unclear risk (2/7).

Liaw 2011.

Study characteristics
Methods	Study design: between participants Total study duration: 2007 to 2008
Participants	Total number: 110 Setting: neonatal nursery at a medical centre in Taipei Diagnostic criteria: Inclusion criteria: 1) gestational age (GA) greater than or equal to 36 weeks; 2) birth weight greater than or equal to 2200 g; 3) Apgar score greater than or equal to 7 at the first and fifth minute after birth; 4) underwent IM injection for hepatitis B; and 5) mothers healthy without substance abuse. Exclusion criteria: 1) congenital anomalies, 2) neurological impairment, 3) documented congenital or nosocomial sepsis; 4) surgery; 5) substance‐abusing mother; 6) had been administered sedatives, analgesics, or naloxone; and 7) no need for injection against hepatitis B Age Minimum: 36.4 weeks Maximum: 42.3 weeks Mean: control: 39.2, non‐nutritive sucking: 39.2, sucrose: 39.2 SD: control: 1.3, NNS: 1.3, sucrose: 1.1 Sex Males: 84 Females: 81 Country: Taiwan
Interventions	Total number of intervention groups: 2 Control group description: during injection procedures, infants in the routine care group (control group) received gentle touch and verbal comfort Total number of participants randomised to each of the groups: 55 Specific intervention #1: non‐nutritive sucking (NNS) Intervention details: infants in this group were given a standard silicone newborn pacifier manufactured with natural scent to stimulate sucking 2 minutes before touching the thigh to initiate the injection procedures Specific intervention #2: sucrose Intervention details: infants in the sucrose group were fed 2 mL of 20% sucrose through a syringe 2 minutes before the injection procedures Integrity of intervention: good
Outcomes	Pain outcomes, time points when measured, and scale limits: NFCS Time points: measured every minute during baseline (T1) and injection procedures (T2), and during recovery at the first (T3), second (T4), third (T5), fourth (T6), and fifth (T7) minutes Upper limit: 48 Lower limit: 0 High score = more intense pain Continuous Cry duration Time points: seconds from the video recording using a stopwatch from the start of the injection procedures to the fifth minute of recovery Greater/longer cry = greater distress Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: pain caused by intra‐muscular injection procedures, as brief as 1 minute, was lower after administering NNS or 20% oral sucrose than that after routine care. Using NNS and sucrose also enhanced infants’ physiological stability and shortened their cry duration during injections. Furthermore, administering sucrose 2 minutes before the injection more effectively reduced newborns’ pain during injection than NNS. Clinicians should incorporate NNS and sucrose into care‐giving while newborns undergo painful procedures. These study results add to evidence supporting the use of NNS and sucrose for pain relief in newborns during immunisation injection. Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	High risk	Personnel not totally blinded but suspect could be guessed based on sucking.
3b) Detection bias (blinding of outcome assessors)	High risk	Coders not completely blinded.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (2/7), unclear risk (0/7).

Liaw 2012.

Study characteristics
Methods	Study design: cross‐over Total study duration: infants were recruited from January 2006 to October 2008
Participants	Total number: 34 Setting: level III neonatal intensive care unit in Taipai Diagnostic criteria: Inclusion criteria: (1) GA 29 to 37 weeks and postmenstrual age 30 to 38.5, (2) post‐birth age 3 to 28 days, and (3) stable condition (score < 12 on the Neonatal Therapeutic Intervention Scoring System [NTISS] for disease severity) Exclusion criteria: (1) congenital anomalies, (2) neurologic impairment including convulsion, intraventricular haemorrhage > grade II or periventricular leukomalacia, (3) documented congenital or nosocomial sepsis, (4) surgery, (5) severe growth restriction at birth, (6) substance‐abusing mother, and (7) severe medical conditions requiring treatments such as sedatives, muscle relaxants, antiepileptic, or analgesic drugs Age: Minimum: actual minimum and maximum ages not specified, only the ranges that were part of the inclusion criteria were specified (i.e. GA of 28 to 37 weeks, postmenstrual age of 30 to 38.5 weeks, and post‐birth age of 3 to 28 days) Maximum: see above under 'minimum' Mean: mean GA of 33.1 weeks; mean postmenstrual age of 34.0 weeks; mean age of 7.0 days SD: SD for GA was 2.0 weeks; SD for postmenstrual age 2.0 weeks; SD for age of 5.0 days Sex Males: 18 (53 % of 34) Females: 16 (47% of 34) Country: Taiwan
Interventions	Total number of intervention groups: 2 Control group description: routine care: infants were given only gentle touch and verbal comfort Total number of participants randomised to each of the groups: intervention #1 = 34; intervention #2 = 34; control = 34 Specific intervention #1: non‐nutritive sucking Intervention details: infants were given a standard silicone newborn pacifier to stimulate sucking while touching the foot to initiate heel stick procedures Specific intervention #2: facilitated tucking Intervention details: infants were in a flexed posture and gently held by a nurse’s warm hands without strongly restraining the infant’s body and limb movements, one hand on the infant’s head, and the other on the buttocks. Facilitated tucking was also provided while the physician assistant touched the infant for heel stick. Integrity of intervention: good
Outcomes	Pain outcomes, time points when measured, and scale limits: PIPP Time points: all measures were standardised from baseline to the 10th minute after heel stick procedures. Upper limit: 21 Lower limit: 0 High score = more pain Continuous Modified preterm infant behavioural coding scheme Time points: all measures were standardised from baseline to the 10th minute after heel stick procedures. The RA coded the behaviours that occurred in every 10‐second interval during the 5 sessions of heel stick procedures. Data were transformed to number of 10‐second intervals in which the behaviour occurred. Higher frequency of behaviour = greater distress Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: both non‐nutritive sucking and facilitated tucking effectively reduced pain scores more than routine care during heel stick procedures. Non‐nutritive sucking reduced PIPP pain scores more effectively than facilitated tucking. However, facilitated tucking showed broader effects not only on relieving pain, but also on enhancing infants’ physiological and behavioural stability during heel stick procedures. Study funding source: National Science Council of Taiwan, ROC for grant support (NSC 95‐2314–B‐016‐057 and NSC 96‐2314–B‐016‐004) Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Unclear risk	Lack of information.
3a) Performance bias (blinding of personnel)	High risk	Personnel not totally blinded.
3b) Detection bias (blinding of outcome assessors)	High risk	Coders not blinded.
4) Attrition bias (incomplete outcome data)	Low risk	A reasonable account of how attrition was dealt with is given.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (2/7), unclear risk (1/7).

Liaw 2013.

Study characteristics
Methods	Study design: between participants Total study duration: infants were screened from August 2010 until December 2011
Participants	Total number: 110 (89 included in quantitative analysis) Setting: level III neonatal intensive care unit (NICU) and a neonatal special care unit at a medical centre in Taipei Diagnostic criteria: Inclusion criteria: (1) gestational age (GA) 26 to 37 weeks and postmenstrual age (PMA) 26.4 to 38 weeks, (2) post‐birth age 2 to 28 days, and (3) disease condition acceptable for observation (illness severity indicated by the Neonatal Therapeutic Intervention Scoring System (NTISS) score ≤ 22) Exclusion criteria: (1) congenital anomalies, (2) neurologic impairment including convulsion, intraventricular haemorrhage > grade II or periventricular leukomalacia, (3) documented congenital or nosocomial sepsis, (4) surgery, (5) severe growth restriction at birth, (6) substance‐abusing mother, and (7) severe medical conditions requiring treatments such as sedatives, muscle relaxants, antiepileptic, or analgesic drugs Age: Minimum: no minimum age stated but GA range is 11   Maximum: no maximum age stated but GA range is 11 Mean: GA = 32.4 weeks, PMA = 33.3 weeks, age = 6.5 days SD: GA = 3.1 weeks, PMA = 3.1 weeks, age = 6.6 days   Sex: Males: 54 (49.5%) Females: 56 (50.5%) Country: Taiwan
Interventions	Total number of intervention groups: 4 Control group description: routine care: infants were given only gentle touch and verbal comfort if they were crying Total number of participants randomised to each of the groups: intervention #1 = 22; intervention #2 = 21; intervention #3 = 21; intervention #4 = 23; control = 23 Intervention descriptions: Non‐nutritive sucking (NNS): infants were given a standard silicone newborn pacifier to stimulate sucking 1 minute before touching the foot to initiate heel stick procedures Oral sucrose: infants were fed 0.2 mL to 2.0 mL of 20% sucrose through a syringe 2 minutes before the heel stick procedures depending on the infants’ GA (GA 26 to 28 weeks: 0.2 mL; GA 28.1 to 30 weeks: 0.5 mL; GA 30.1 to 32 weeks: 1 mL; GA 32.1 to 37 weeks: 1.5 mL; GA > 37 weeks: 2.0 mL) Facilitated tucking (FT): infants were in a flexed posture and gently held by the intervener’s warm hands without strongly restraining the infant’s head and body, one hand on the infant’s head, and the other on the trunk. FT was provided while the infant was touched by the senior nurse performing heel sticks Specific intervention #1: NNS‐FT group Specific intervention #2: FT‐sucrose group Specific intervention #3: NNS‐sucrose group Specific intervention #4: NNS‐sucrose‐FT group Integrity of intervention: satisfactory
Outcomes	Pain outcomes, time points when measured, and scale limits: Active sleep Time points: data collection was conducted at 1‐minute intervals. All data were collected over 4 phases: (a) baseline (P1, 30 minutes without stimuli before heel stick procedures); (b) intervention (P2, one of 5 treatment conditions), (c) heel stick procedures (P3), and (d) recovery (P4, 30 minutes after P3 from when the research nurse finished collecting blood and left the infant to the 30th minute).   Range of possible scores: not reported Transition state Time points: data collection was conducted at 1‐minute intervals. All data were collected over 4 phases: (a) baseline (P1, 30 minutes without stimuli before heel stick procedures); (b) intervention (P2, one of 5 treatment conditions), (c) heel stick procedures (P3), and (d) recovery (P4, 30 minutes after P3 from when the research nurse finished collecting blood and left the infant to the 30th minute).   Range of possible scores: not reported Active awake and quiet awake Time points: data collection was conducted at 1‐minute intervals. All data were collected over 4 phases: (a) baseline (P1, 30 minutes without stimuli before heel stick procedures); (b) intervention (P2, one of 5 treatment conditions), (c) heel stick procedures (P3), and (d) recovery (P4, 30 minutes after P3 from when the research nurse finished collecting blood and left the infant to the 30th minute).   Range of possible scores: not reported Fussing or crying Time points: data collection was conducted at 1‐minute intervals. All data were collected over 4 phases: (a) baseline (P1, 30 minutes without stimuli before heel stick procedures); (b) intervention (P2, one of 5 treatment conditions), (c) heel stick procedures (P3), and (d) recovery (P4, 30 minutes after P3 from when the research nurse finished collecting blood and left the infant to the 30th minute).   Range of possible scores: not reported Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: the 4 treatment combinations differentially reduced infants’ high arousal across heel stick procedures. The combined use of oral sucrose‐tucking, sucking‐oral sucrose, and sucking‐oral sucrose‐tucking more effectively reduced occurrences of infant fussing or crying than routine care. Treatment combinations of sucking‐oral sucrose‐tucking and sucking‐oral sucrose also better facilitated infants’ sleep than routine care. To preserve infants’ sleep, clinicians should use combinations of non‐nutritive sucking, oral sucrose, and facilitated tucking to reduce agitation during painful procedures. Study funding source: the work was supported by a grant from National Science Council of Taiwan, ROC (NSC 95‐2314‐B‐016‐057 and NSC 96‐2314‐B‐016‐004) Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Unclear risk	Lack of information.
3a) Performance bias (blinding of personnel)	High risk	Personnel not totally blinded.
3b) Detection bias (blinding of outcome assessors)	High risk	Coders not blinded.
4) Attrition bias (incomplete outcome data)	Low risk	A reasonable account of how attrition was dealt with is given.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (3/7), high risk (2/7), unclear risk (2/7).

Liu 2010.

Study characteristics
Methods	Study design: randomised controlled trial, between‐group (parallel‐group controlled trial) Total study duration: 5 months (June to October 2006)
Participants	Total number: 70 Setting: newborn special care nursery and in a newborn baby room of a teaching hospital in northern Taiwan Diagnostic criteria: neonates were greater than or equal to 32 weeks gestational age, medically stable, scheduled to undergo a newborn screening procedure within 24 hours to 7 days, Apgar scores greater than 7 at 1 and 5 minutes, with no crying in the 5 minutes pre‐venipuncture Age: Minimum: unknown Maximum: unknown Mean: 3.0 days SD: 1.1 days Sex: Males: 48 Females: 57 Country: Taiwan
Interventions	Total intervention groups: 1 Control group description: no intervention provided prior to venipuncture. Infants were taken to a quiet, individual room and placed on a heated radial warmer, naked except for their diaper. Infants were observed for 2 minutes to collect baseline data. Total number of participants randomised to each of the groups: 35 in each group  Specific intervention: non‐nutritive sucking Intervention details: each neonate was assisted to suck on a pacifier for 2 minutes pre‐sterilisation. The pacifier was continuously provided until 2 minutes post‐needle. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured: NIPS Time points: coded at 1‐minute intervals for 2 minutes in preparation, venipuncture, and recovery phases Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: non‐nutritive sucking can effectively decrease the level of pain Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes the potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	A convincing effort to reduce bias in outcomes measurement is reported.
3b) Detection bias (blinding of outcome assessors)	Low risk	Coders were blinded.
4) Attrition bias (incomplete outcome data)	High risk	There is no documented evidence or insufficient evidence reported of how attrition was dealt with.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (1/7), unclear risk (1/7).

Lopez 2015.

Study characteristics
Methods	Study design: between‐group Total study duration: 4 months
Participants	Total number: 42 Setting: hospital (NICU) Diagnostic criteria: preterm neonate Gestational age: Mean: 30.6 (facilitated tucking), 31.7 (control) SD: 2.8 (facilitated tucking), 1.9 (control) Postnatal age: Mean: 3.2 days (facilitated tucking), 4.5 days (control) SD: 3.7 (facilitated tucking), 7.1 (control) Sex: Males: 24 Females: 18 Country: Malaysia
Interventions	Total number of intervention groups: 1 Control group description: normal NICU routine Total number of participants randomised to each of the groups: Facilitated tucking: 21 Control: 21 Specific intervention: facilitated tucking Intervention details: the treatment group was exposed to venepuncture when held in a tucked position. The baby is held for 2 minutes before the procedure until 30 seconds after. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure and 1 minute post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of the study authors: the findings of this study suggest that facilitated tucking is able to alleviate pain Study funding source: The Health and Translational Cluster (Grant Number: RG 396‐11HTM and UMRG RP010‐2012E), the Faculty of Medicine, University of Malaya Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	While authors describe that the envelopes were sealed, there is no information about whether the envelopes were sequentially numbered or opaque. There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	There is no mention of blinding of study personnel, but it would not be possible to blind study personnel to treatment condition given the covert nature of the intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors were not blinded.
4) Attrition bias (incomplete outcome data)	Low risk	Reasons for missing outcome data unlikely to be related to true outcome (for survival data, censoring unlikely to be introducing bias).
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain is "high risk" then the overall rating is high risk.

Malakian 2017.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 64 Setting: hospital (NICU) Diagnostic criteria: full‐term neonates Gestational age: Intervention: 36% 37 weeks GA; 64% 38 weeks GA Control: 48% 37 weeks GA; 52% 38 weeks GA Sex: Males: 39 Females: 25 Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: in the control group, a 1 × 1 cm room‐temperature pack was placed on the side of infants' foot (blood drainage area) for 20 seconds Total number of participants randomised to each of the groups: Intervention: 32 Control: 32 Specific intervention: cold pack Intervention details: a 1 cm × 1 cm ice pack was placed on the side of infant’s foot (blood drainage area) for 20 seconds Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured CRIES Time points: within 30 seconds of the painful procedure Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: this study showed that local cold can reduce the level of pain caused by heel prick blood sampling and can decrease early and late complications of pain in infants Study funding source: financial support from the Jundishapur University of Medical Sciences Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Unable to blind personnel due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessor blinded to nature of intervention.
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data observed.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, an overall score of 'high' is given when one or more domain is determined to be high.

McGinnis 2016.

Study characteristics
Methods	Study design: between‐group Total study duration: 9 months
Participants	Total number: 56 Setting: hospital (NICU) Diagnostic criteria: full‐term neonates Gestational age: Median: 36 (sucrose + swaddling), 38 (sucrose + swaddling + vibration) IQR: 22 to 38 (sucrose + swaddling), 30 to 39 (sucrose + swaddling + vibration) Sex: Males: 31 Females: 25 Country: USA
Interventions	Total number of intervention groups: 2 Total number of participants randomised to each of the groups: sucrose + swaddling = 26, sucrose + swaddling + vibration = 30 Specific intervention #1: sucrose + swaddling Intervention details: patients randomised to the control group were first swaddled with one leg exposed and a heel warmer placed on the exposed heel 3 to 5 minutes before the lance. 2 minutes before the heel stick, the patient was given 0.1 mL to 0.2 mL of oral 24% sucrose with pacifier offered. Before the lance, the heel warmer was removed and site prepared with alcohol. Using the appropriate‐sized Tenderfoot lancet, the outer edge of the lateral plantar surface was lanced, along the sural dermatome. A period of 10 seconds with‐out squeezing was observed to estimate the response to the heel stick alone. After this undisturbed phase, blood was collected with intermittent squeezing and allowing for capillary refill. After 0.4 mL to 1 mL of capillary blood was collected, a small adhesive bandage was applied to the heel and the infant was re‐swaddled in a flexed, midline posture. Integrity of intervention #1: good Specific intervention #2: sucrose + swaddling + vibration Intervention details #2: the experimental group procedure was identical to the control except for the following: approximately 30 seconds before the heel stick, the vibrator was applied over the mid/lateral calf, just below the knee in accordance with the sural dermatome for a 30‐second “test” vibration. During this period, the infant was observed carefully for any adverse responses. If the infant had any apnoea, bradycardia, or desaturation, the vibrator was immediately removed and reaction documented. Similarly, if the infant demonstrated sustained heart rate elevation 20% above baseline or persistent crying, the mechanical vibration was immediately removed and not reapplied. If there were no adverse responses observed during the 30‐second trial, the vibration was continued throughout the duration of the heel lance procedure and was stopped immediately before the application of the adhesive bandage. Adverse responses were recorded. The site of applied vibration was monitored for any redness, swelling, bruising, or alteration in skin integrity. Integrity of intervention #2: good
Outcomes	Pain outcomes, scale limits, and time points when measured N‐PASS Time points: first 30 seconds post‐procedure and 2 minutes post‐procedure Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of the study authors: applied mechanical vibration is a safe and effective method for managing heel lance pain Study funding source: Dudley L. Moore Nursing and Allied Health Fund Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Unable to blind personnel due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	Blinding was not possible because outcome was assessed at time of heel lance (i.e. not video recorded for later assessment) and study group would have been obvious.
4) Attrition bias (incomplete outcome data)	Low risk	Reasons for missing outcome data unlikely to be related to true outcome (for survival data, censoring unlikely to be introducing bias); missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, an overall score of 'high' is given when one or more domain is determined to be high.

Mir 2018.

Study characteristics
Methods	Study design: between‐group Total study duration: 2 months
Participants	Total number: 78  Setting: clinic Diagnostic criteria: healthy full‐term neonates Postnatal age: Mean: 4.8 days (Yakson), 4.1 days (heel warming), 5.1 days (control) SD: 1.0 (Yakson), 0.7 (heel warming), 2.4 (control) Sex: Males: 39 Females: 39 Country: Iran
Interventions	Total number of intervention groups: 2 Description of control: the participants in the control group were placed in a supine position lying without any interventions before taking blood samples from their heels and then heel stick blood sampling was conducted Total number of participants randomised to each of the groups: Yakson = 26, heel warming = 26, control = 26 Specific intervention #1: Yakson Intervention details #1: in the first intervention group, the non‐dominant hand was placed on the neonates’ back and the dominant hand was put on their chest and abdomen. According to the protocol of Yakson touch, the researcher wore a clean gown and washed their hands thoroughly with antimicrobial agents for 3 minutes and then warmed their hands using a warmer up to the temperature of 34°C. The temperature of the hand was measured using a digital thermometer. At the next step, the researcher relaxed their both arms and shoulder muscles for a minute and took a deep breath to concentrate the vital energy in the palm of their hands. Then, the researcher started Yakson therapeutic touch as explained for 15 minutes. In addition, 5‐minute was considered for the palm rest, while one hand was resting on the neonate’s chest and abdomen, and the neonate was supported with the other hand. Additionally, the researcher concentrated the energy in their resting hands with a deep breath to bring a healthy vital energy to the neonate. During this interval, the researcher breathed slowly to keep their calm; thereafter, they had gentle touches for 5 minutes in the same hand position and repeated the rest for 5 minutes, touched the neonate (one minute), had a rest (30 seconds), touched again (1 minute), rested (30 seconds), and touched for the third time (2 minutes). Moreover, the researcher touched the neonate’s abdomen and chest in a clockwise direction and circular motion with the diameter of 1 cm every 10 seconds with 5 minutes of rest again and repeated their hand rest in the same way as described above. During Yakson therapeutic touch, the palm and all the fingers of the researcher were constantly in close contact with the neonate; thus, the newborn felt no pressure Integrity of intervention #1: good Specific intervention #2: heel warming Intervention details #2: water at 40°C was poured into a bottle, and then the bottle was placed for 5 minutes on the opposite direction of the hole on the outside of the neonate’s heel. Moreover, heel stick procedure was performed immediately after the removal of the hot‐water bottle Integrity of intervention #2: good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time point: first 30 seconds post‐procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of the study authors: the use of Yakson therapeutic touch could relieve pain, soothe the neonates, and shorten cry duration in newborns after heel stick procedure. Nevertheless, heel warming only increased blood flow for easier blood sampling. Study funding source: Mashhad University of Medical Sciences, Mashhad, Iran Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel were stated not to be blinded.
3b) Detection bias (blinding of outcome assessors)	High risk	The researcher, who was stated to be un‐blinded, was the outcome assessor.
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data observed.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall assessment is high risk.

Mirzarahimi 2013.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 90 Setting: NICU in Alavi hospital at Ardebil Diagnostic criteria: healthy full‐term neonates Age: Minimum: 3 days Maximum: 7 days Mean: N/A SD: N/A Sex: Males: 54 Females: 36 Country: Iran
Interventions	Total number of intervention groups: 2 Control group description: "received no intervention" Total number of participants randomised to each of the groups: intervention #1 = 30; intervention # 2 = 30; control = 30 Intervention descriptions: Specific intervention #1: NNS – researcher gave each neonate a standard small, short, hollow soft latex nipple while applying gentle pressure to maintain it in the neonate’s mouth Specific intervention #2: massage – 2 minutes before the heel stick, one of the investigators slowly massaged the outer aspect of the leg chosen for the heel stick from toes to mid thigh by using a firm but gentle pressure by fingers and thumbs. The massage pressure was adjusted to keep the baby comfortable. At the end of the massage intervention (2 minutes), the heel was wrapped in a warm cloth for one minute. An assistant nurse performed the heel stick using a lancet. Integrity of intervention: good
Outcomes	Pain outcomes, time points when measured, and scale limits: PIPP Time points: did not indicate timing of pain outcome, but seems it was during heel stick Range of possible scores: 0 (no pain) to 21 (maximum pain) – maximum value in this study was 18 (still reflecting greater pain) Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: single or combined use of massage and a pacifier is effective in reducing pain response in neonates undergoing heel stick Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Insufficient information.
2) Selection bias (allocation concealment)	Unclear risk	Insufficient information.
3a) Performance bias (blinding of personnel)	Unclear risk	Insufficient information.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	Insufficient information.
4) Attrition bias (incomplete outcome data)	Low risk	No incomplete outcome data, p.77.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (2/7), high risk (0/7), unclear risk (5/7).

Mitchell 2016.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 153  Setting: hospital Diagnostic criteria: healthy neonates between 37 to 42 weeks Gestational age: Mean: 39.0 weeks (overall) SD: 1.2 (overall) Sex: Males: not reported Females: not reported Country: USA
Interventions	Total number of intervention groups: 4 Total number of participants randomised to each of the groups: NESAP + sucrose = 40, sham NESAP + sucrose = 37, sham NESAP + water = 39, NESAP + water = 37 Specific intervention #1: NESAP + sucrose Intervention details: all treatment groups received facilitated tucking and a pacifier. Facilitated tucking was administered by a trained advanced practice nurse researcher and was applied by providing gentle support to the head and flexing the knees to midline. Small electrodes were placed on the baby’s legs at specific acupuncture points: ZuSanLi (ST36), SanYinJiao (SP6), KunLun(Bl60) and TaiXi (KI3). StimCare electrodes with a gel base were applied; electrodes were wrapped with gauze to ensure that placement remained intact during the heel stick procedure. The TENS unit was turned on, and settings were increased very gradually over 60 seconds to avoid startling the infant. A low continuous current of 3.5 mA and frequency of 10 Hz were delivered for 10 ± 1 minute prior to the heel stick, with continued stimulation during and for 2 minutes after the heel stick. Newborns received 1 mL ± 0.1 mL of the 24% sucrose solution orally with pacifier at 2 ± 0.5 minutes prior to the procedure. Integrity of intervention #1: good Specific intervention #2: sham NESAP + sucrose Intervention details #2: all treatment groups received facilitated tucking and a pacifier. Facilitated tucking was administered by a trained advanced practice nurse researcher and was applied by providing gentle support to the head and flexing the knees to midline. Sham stimulation was performed by applying electrodes adjacent to the proposed acupuncture points (ST36, SP6, KI3, BL60), but the TENS unit was not switched on. Newborns received 1 mL ± 0.1 mL of 24% sucrose solution orally with pacifier at 2 ± 0.5 minutes prior to the procedure. Integrity of intervention #2: good Specific intervention #3: sham NESAP + water Intervention details #3: all treatment groups received facilitated tucking and a pacifier. Facilitated tucking was administered by a trained advanced practice nurse researcher and was applied by providing gentle support to the head and flexing the knees to midline. Sham stimulation was performed by applying electrodes adjacent to the proposed acupuncture points (ST36, SP6, KI3, BL60), but the TENS unit was not switched on. Newborns received 1 mL ± 0.1 mL of sterile water with pacifier at 2 ± 0.5 minutes prior to the procedure. Integrity of intervention #3: good Specific intervention #4: NESAP + water Intervention details #4: all treatment groups received facilitated tucking and a pacifier. Facilitated tucking was administered by a trained advanced practice nurse researcher and was applied by providing gentle support to the head and flexing the knees to midline. Small electrodes were placed on the baby’s legs at specific acupuncture points: ZuSanLi (ST36), SanYinJiao (SP6), KunLun(Bl60) and TaiXi (KI3). StimCare electrodes with a gel base were applied; electrodes were wrapped with gauze to ensure that placement remained intact during the heel stick procedure. The TENS unit was turned on, and settings were increased very gradually over 60 seconds to avoid startling the infant. A low continuous current of 3.5 mA and frequency of 10 Hz were delivered for 10 ± 1 minutes prior to the heel stick, with continued stimulation during and for 2 minutes after the heel stick. Newborns received 1 ± 0.1 mL of sterile water with pacifier at 2 ± 0.5 minutes prior to the procedure. Integrity of intervention #4: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure and 2 minutes post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of the study authors: NESAP at 3.5 mA, 10 Hz is not effective in relieving pain during heel sticks in neonates Study funding source: Mayday fund and NIGMS Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	Low risk	Authors state that research staff was blinded to study condition, and because the TENS unit was silent and both groups received a substance by pacifier, personnel could be blinded to which infants were receiving NESAP (vs sham NESAP) and sucrose (vs water).
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were blinded.
4) Attrition bias (incomplete outcome data)	Low risk	Reasons for missing outcome data were described and were unlikely to be related to true outcome. Missing outcome data balanced in numbers across groups, with similar reasons for missing data across groups.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, a rating of "unclear" in one or more domain leads to an overall rating of unclear.

Morrow 2010.

Study characteristics
Methods	Study design: between‐group Total study duration: 3 months (February to April 2008)
Participants	Total number: 42 Setting: tertiary hospital serving a major metropolitan area Diagnostic criteria: infants scoring in the high‐intermediate or high‐risk zone by TcB (transcutaneous bilirubin) measurement receiving a total serum bilirubin (TSB) evaluation; full‐term neonates Age: greater than or equal to 37 weeks gestation Minimum: not reported Maximum: not reported Mean: 39.1 weeks gestation SD: 1.8 days Sex: Males: 18 Females: 21 Note: The authors reported that 42 participants were included but only reported on sex for 39 participants Country: USA
Interventions	Total intervention groups: 1 Control group description: infants were placed in a standard position (supine while lying in a crib) during a sample collection. The crib was elevated to a 30 degree angle and the leg was elevated during the heel lance. Total number of participants randomised to each of the groups: Intervention: 22 Control: 20 Specific intervention: swaddling and holding upright Intervention details: infants were swaddled and held upright at a 90 degree angle, with one leg exposed Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured: NIPS Coded immediately after heel lance Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: swaddling infants while holding them in an upright position was superior for pain relief during heel lance procedures when compared with a standard position technique Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method of randomisation is reported but this did not involve an independent person.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	High risk	Nurses were not blinded.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	Not specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (3/7), high risk (2/7), unclear risk (2/7).

Najafi 2019.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 113 Setting: hospital Diagnostic criteria: full‐term healthy neonates Age: Average age between 5.6 to 14.4 hours Sex: Males: 56.4% Females: 43.6% Country: Iran
Interventions	Total number of intervention groups: 2 Control group description: injection was done by the typical hospital procedures. Total number of participants randomised to each of the groups: cooling gel pad = 39, non‐nutritive sucking = 35, control = 39 Specific intervention: cooling gel pad, non‐nutritive sucking Intervention details: cooling gel pad: the pad is frozen for 3 hours and placed on the right thigh of the neonates. They did not place the pad directly on the skin rather they put down a gauze directly on the skin and placed the pad on the gauze. They placed the pad indirectly for 90 seconds before the injection. Before and after the cooling gel pad was placed, the body temperature of the neonate was measured using a digital thermometer (placed in the ear) to check for potential hypothermia. Immediately after the cooling gel pad was removed, the injection happened. Non‐nutritive sucking: The researcher cleaned their hands and provided their little finger for non‐nutritive sucking for about 2 minutes. After 2 minutes while the neonate was sucking the finger, the injection happened. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: within first 30 seconds post‐procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: both methods of local cold and non‐nutritive sucking are helpful in reducing pain intensity. This intervention is practical and easy, which could minimise the negative impact of pain in neonates. Study funding source: not mentioned Author COI statement: not mentioned
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Low risk	A convincing method for allocation concealment is reported.
3a) Performance bias (blinding of personnel)	Unclear risk	It is not clear whether the personnel were blinded to the intervention.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were blinded to study condition.
4) Attrition bias (incomplete outcome data)	Unclear risk	Attrition/exclusions not described so unable to determine whether there was incomplete outcome data.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, if one or more domain is "unclear", the overall judgement is unclear.

Ozkan 2019.

Study characteristics
Methods	Study design: between‐group Total study duration: 11 months
Participants	Total number: 139 Setting: hospital Diagnostic criteria: full‐term healthy Gestational age (weeks): Mean: 38.7 (acupressure), 38.7 (foot massage), 38.6 (control), overall mean = 38.7 SD: 0.7 (acupressure), 0.9 (foot massage), 0.8 (control), overall SD = 0.8   Sex: Males: 27 (acupressure), 20 (foot massage), 26 (control), total = 73 Females: 19 (acupressure), 27 (foot massage), 20 (control), total = 66 Country: Turkey
Interventions	Total number of intervention groups: 2 Control group description: no description Total number of participants randomised to each of the groups: acupressure = 46; foot massage = 47; control = 46 Specific intervention: acupressure, foot massage Intervention details: acupressure: for 2 minutes, acupressure was performed in the acupressure points (Kun Lun (UB60) and Taixi (K3). The acupuncture points KunLun (UB60) and Taixi (K3) are on the side of the ankle. These points are effective in reducing pain. Each point was applied acupressure for 60 seconds, and heel lancing was performed right after this procedure. Foot massage: neonates were given foot massage for 2 minutes, and heel lancing was performed right after the massage. Integrity of intervention: acupressure = good, foot massage = satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: during heel lance (0 minutes), 1 minute after heel lance Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: acupressure and massage were effective in reducing pain in neonates during the heel lancing procedure. However, there was no statistically significant difference between acupressure and massage. Acupressure and massage techniques can be applied for procedural pain management in the newborn. Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel unable to be blinded due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors unable to be blinded due to nature of outcome (leg movements coded) and intervention.
4) Attrition bias (incomplete outcome data)	Low risk	Showed attrition reasons which appear balanced across groups.
5) Selective reporting (reporting bias)	Unclear risk	While a trial register was reported, the study was retroactively registered so unable to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Peng 2018.

Study characteristics
Methods	Study design: between‐group Total study duration: 12 months
Participants	Total number: 74 Setting: hospital (NICU) Diagnostic criteria: preterm infants Gestational age: Mean: 31.3 (NNS + breast milk), 31.3 (NNS + breast milk + facilitated tucking) SD: 2.8 (NNS + breast milk), 2.9 (NNS + breast milk + facilitated tucking) Postnatal age: Mean: 13.8 days (NNS + breast milk), 12.7 days (NNS + breast milk + facilitated tucking) SD: 12.0 (NNS + breast milk), 10.1 (NNS + breast milk + facilitated tucking) Sex: Males: not reported Females: not reported Country: Taiwan
Interventions	Total number of intervention groups: 2 Total number of participants randomised to each of the groups: NNS + breast milk = 37, NNS + breast milk + facilitated tucking = 37 Specific intervention #1: NNS + breast milk Intervention details #1: for the sucking condition, infants received a standard silicone newborn pacifier 2 minutes before their foot was touched to start heel stick procedures. For the breast milk condition, infants were orally fed 0.5 mL to 2.0 mL of expressed breast milk through a syringe 2 minutes before heel stick. The volume of breast milk fed to the infant was based on gestational age (gestational age 27 to 28 weeks: 0.5 mL; gestational age 28.1 to 30 weeks: 1.0 mL; gestational age 30.1 to 32 weeks: 1.5 mL; gestational age 32.1 to 37 weeks: 2 mL). Integrity of intervention #1: poor Specific intervention #2: NNS + breast milk + facilitated tucking Intervention details #2: for the sucking condition, infants received a standard silicone newborn pacifier 2 minutes before their foot was touched to start heel stick procedures. For the breast milk condition, infants were orally fed 0.5 mL to 2.0 mL of expressed breast milk through a syringe 2 minutes before heel stick. The volume of breast milk fed to the infant was based on gestational age (gestational age 27 to 28 weeks: 0.5 mL; gestational age 28.1 to 30 weeks: 1.0 mL; gestational age 30.1 to 32 weeks: 1.5 mL; gestational age 32.1 to 37 weeks: 2 mL). For the tucking condition, infants were gently held in warm hands before the heel stick, keeping the infant in a flexed, mid‐line position with his/her 4 limbs close to his/her own body and minimum restraint of the infant’s head and body. Integrity of intervention #2: poor
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure and 1 minute post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of the study authors: the combined use of sucking + breast milk + tucking and sucking + breast milk effectively reduced preterm infants’ mild pain and moderate‐to‐severe pain during heel stick procedures. Adding facilitated tucking helped infants recover from pain across 8 phases of heel stick procedures. Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	No mention of blinding of study personnel.
3b) Detection bias (blinding of outcome assessors)	High risk	Unable to blind assessors to NNS condition given the pacifier.
4) Attrition bias (incomplete outcome data)	Low risk	109 infants that met eligibility criteria and who were receiving heel sticks, and 109 infants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall assessment is high risk.

Perroteau 2018.

Study characteristics
Methods	Study design: between‐group Total study duration: 17 months
Participants	Total number: 59 Setting: hospital Diagnostic criteria: preterm infants Gestational age: Median: 29 (NNS), 30 (NNS + facilitated tucking), 29 (overall) IQR: 28 to 31 (NNS), 29 to 30 (NNS + facilitated tucking), 28 to 31 (overall) Sex: Males: 28 Females: 31 Country: France
Interventions	Total number of intervention groups: 2 Total number of participants randomised to each of the groups: NNS = 29, NNS + facilitated tucking = 30 Specific intervention #1: NNS Intervention details #1: the nurse assistant maintained the Evoprene pacifier for 3 minutes after the heel stick procedure Integrity of intervention #1: good Specific intervention #2: NNS + facilitated tucking Intervention details #2: either a nurse or nurse assistant prepared the infant in the facilitated tucking position at least 15 seconds before the heel stick procedure. The infant's arms and legs were held by the nurse inflexed positions close to the midline of the chest, while one nurse’s hand was gently placed under the infant’s neck and maintained the pacifier with sterile water; the other hand was placed on the buttocks. The facilitated tucking procedure was maintained for 3 minutes after the heel stick procedure. The nurse assistant maintained the Evoprene pacifier for 3 minutes after the heel stick procedure. Integrity of intervention #2: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of the study authors: the PIPP pain scores did not differ between the intervention group and the control group Study funding source: European Science Foundation Research Network Programme Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Due to nature of intervention, personnel could not be blinded to study condition.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were blinded to study condition.
4) Attrition bias (incomplete outcome data)	Low risk	Reason for missing outcome data explained and unlikely to be related to true outcome.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, because one domain received a rating of "high risk", the overall rating is high.

Petrillo 2010.

Study characteristics
Methods	Study design: between‐group Total study duration: 3 weeks
Participants	Total number: 85 Setting: community paediatric clinic Diagnostic criteria: healthy preterms and term neonates   Gestational age: 32 to 41 Mean: not mentioned  SD: not mentioned  Sex: Males: not mentioned  Females: not mentioned  Country: Italy
Interventions	Total number of intervention groups: 2 Control group description: term and premature babies are treated with 20% OS. For the execution of the procedure, a quiet room was identified for full‐term newborns and an acoustically insulated box for premature babies. The sampling procedure involved: short massage of the limb chosen to facilitate the flow of blood to the foot, disinfection of the heel, puncture with a disposable sterile automatic lancet (Tenderfoot), squeezing of the heel and finally collection with a capillary tube.  Total number of participants randomised to each of the groups: Group A (oral sucrose only): 20 neonates and 5 preterms  Group B (oral sucrose and sensorial saturation): 20 neonates and 5 preterms  Group C (sensorial saturation with the presence of mother): 25 neonates     Specific intervention #1: sensorial saturation (SS) without the presence of the mother   Intervention details #1: term and premature babies are treated with SS without the presence of the mother. SS intervention consisted of: use of the pacifier, containment, sweet tone of voice, in addition to the use of OS. For the execution of the procedure, a quiet room was identified for full‐term newborns and an acoustically insulated box for premature babies. The sampling procedure involved: short massage of the limb chosen to facilitate the flow of blood to the foot, disinfection of the heel, puncture with a disposable sterile automatic lancet (Tenderfoot), squeezing of the heel and finally collection with a capillary tube.  Integrity of intervention #1: poor Intervention details #2: sensorial saturation (SS) with presence of mother Intervention details #2: only term babies are treated with SS with the presence of the mother. SS intervention consisted of: use of the pacifier, containment, sweet tone of voice, in addition to the use of OS. For the execution of the procedure, a quiet room was identified for full‐term newborns. The sampling procedure involved: short massage of the limb chosen to facilitate the flow of blood to the foot, disinfection of the heel, puncture with a disposable sterile automatic lancet (Tenderfoot), squeezing of the heel and finally collection with a capillary tube.  Integrity of intervention #2: poor
Outcomes	Pain outcomes, scale limits, and time points when measured   NIPS (term babies)  Time points: before the procedure (t0); after the puncture (t1); after squeezing (t2, only for premature babies); 15 minutes from the conclusion of the withdrawal (t3) NIPS upper limit: 7 NIPS lower limit: 0 High score = more pain Continuous   PIPP (premature babies) Time points: before the procedure (t0); after the puncture (t1); after squeezing (t2, only for premature babies); 15 minutes from the conclusion of the withdrawal (t3) PIPP Upper limit: 21 PIPP Lower limit: 0 High score = more pain Continuous   Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors: The premature heel puncture pain has its maximum peak in the squeezing phase. Sensory saturation is more effective than the administration of glucose solution alone and it does not require a significantly longer care time than glucose administration. Heel withdrawal pain is "unnecessary pain" because there are means to effectively counteract it.  Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Authors do not sufficiently describe the method used to generate the randomisation.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel unable to be blinded due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors unable to be blinded due to nature of intervention.
4) Attrition bias (incomplete outcome data)	Unclear risk	Attrition/exclusions not described so unable to determine whether there was incomplete outcome data.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain is "high", the overall judgement is high.

Pillai Riddell 2018.

Study characteristics
Methods	Study design: between‐group Total study duration: 10 months
Participants	Total number: 128 (64 6‐month olds, 64 18‐month olds) Setting: clinic Diagnostic criteria: full‐term healthy infants Postnatal age: Mean: 64 6‐month olds and 64 18‐month olds Sex: Males: N/A Females: N/A Country: Canada
Interventions	Total number of intervention groups: 1 Description of control: the placebo video provided parents with neutral in‐formation (still in an ABCD format) in lieu of the aforementioned ABCDs. In terms of the ABCDs for the placebo video, parents were encouraged and instructed to, in order, (A) Act in their young child’s best interest, (B) Be aware that needles are distressing, (C) Carry out what they think is best, and (D) Do their best to help their young child. Total number of participants randomised to each of the groups: treatment video = 64, placebo video = 64 Specific intervention #1: treatment video Intervention details #1: the treatment video instructed parents on the ABCDs (Assess anxiety, Belly breathe, Calm Close Cuddle, Distraction) of pain management. These 4 areas of focus were based on extensive analyses of natural soothing behaviours, objectively coded as sensitive, that parents demonstrated in the Opportunities to Understand Childhood Hurt (OUCH) cohort longitudinal study. In regards to the specifics of the ABCDs, parents were encouraged and instructed to, in order, (A) Assess their own anxiety by self‐reflecting on their stress level right before the vaccination begins, (B) Belly breathe if stressed, by putting their hand on their diaphragm and breathing in deeply counting to 3 and exhaling slowly counting to 3, (C) Use a calm, close, cuddle (before, during, and after the needle) with their young child, and (D) Distract their baby when the peak distress has passed. Integrity of intervention #1: good
Outcomes	Pain outcomes, scale limits, and time points when measured MBPS Time point: first 30 seconds post‐procedure, 2 minutes post‐procedure Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of the study authors: the ABCD pain management strategy delivered via video was an effective way to reduce toddler pain after vaccination and increase parental use of rocking and physical comforting. The treatment effect was not demonstrated with infants. Study funding source: this work was supported by salary, operating, and infrastructure awards from the Canada Foundation for Innovation (grant number 532009); Canadian Institutes of Health Research (grant number 524563); and the Ontario Ministry of Research and Innovation awarded to RPR and awards to M.C.O. and L.C. from the Canadian Institutes of Health Research. M.C.O. and L.C. are also trainee members of Pain In Child Health (PICH), a strategic research training initiative of the Canadian Institutes of Health Research Author COI statement: Anna Taddio declares research funding from Pfizer (which funded the video from the trial) and study supplies from Ferndale and Natus for a separate trial
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	RA's and healthcare providers were all blinded to study condition.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were blinded to study allocation.
4) Attrition bias (incomplete outcome data)	Low risk	Flow chart provided and it seems that there were no missing data.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Low risk	As per Sterne 2019, given that all domains received an assessment of "low risk" the overall assessment is low risk.

Ranjbar 2020.

Study characteristics
Methods	Study design: cross‐over Total study duration: 19 months
Participants	Total number: 60 Setting: hospital Diagnostic criteria: hospitalised preterm newborns; infants seem stable Age (weeks): Mean: 32.4 SD: 2.8   Sex: Males: 37 Females: 23 Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: 30 minutes before the heel stick procedure infants were placed in a quiet location. The measurement of pain in the control measurements was performed without any pain relief intervention except the application of gentle touching and verbal comfort. Total number of participants randomised to each of the groups: 60 Specific intervention: facilitated tucking Intervention details: 30 minutes before the heel stick procedure infants were placed in a quiet location. The instructed nurses placed the babies in facilitated tucking position by placing them on their side, with their back gently bent, and their legs were in a flexion angle of greater than 90 degrees. The infants’ shoulders were also constricted up to 90 degrees, and the hands of the nurse were placed over the head close to the mouth or on the infant’s face. Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: before, during, and 5 minutes after heel stick Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: facilitated tucking is an effective method of pain reduction Study funding source: Tehran University of Medical Sciences Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	High risk	Due to sequence generation method that was described, researchers would be aware of group assignment and allocation concealment would therefore not be maintained.
3a) Performance bias (blinding of personnel)	High risk	Blinding not possible due to nature of intervention (facilitated tucking).
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessment appears to have been conducted at bedside, so blinding was not possible.
4) Attrition bias (incomplete outcome data)	Low risk	Showed attrition reasons which appear balanced across groups.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Razaghi 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 80 Setting: hospital Diagnostic criteria: full‐term healthy Age (days): Mean: 5.5 (lavender scent), 5.6 (control) SD: 2.0 (lavender scent), 2.4 (control)   Sex: Males: 15 (lavender scent), 19 (control) Females: 25 (lavender scent), 21 (control) Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: the control group did not receive any specific intervention and received routine care Total number of participants randomised to each of the groups: Lavender scent: 40 Control: 40 Specific intervention: lavender scent Intervention details: neonates inhaled 10 drops of lavender essential oil 0.5% (3 drops of lavender essential oil 100 % was dissolved in 30 mL glycerin solution, which resulted in lavender essential oil 0.5%, produced by Barij Essence Co. Iran, Kashan). A sterile gauze pad (10 cm x 10 cm) wetted with 10 drops of the lavender solution was placed in the incubator ∼10 cm from the neonate’s head and was kept until the next morning. Whenever the neonate required breastfeeding or cried, the mother could take the neonate out from the incubator and again put him/her back. Also, it was ensured that the neonate did not come in contact with other aromatic substances when taken out of the incubator. On the next morning of familiarisation, the scented gauze was removed (after an average familiarisation duration of 8 hours) and was again used at the time of blood sampling in this group. Blood sampling was done a half to 1 hour (on average 43.38 minutes) after the removal of the scented gauze. Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured DAN Time points: scores of DAN scale were observed, calculated, and recorded by a trained person. In the first 3 minutes of blood sampling, the duration of crying (in seconds) was measured from start (the first audible sound) to end (silence that continues at least 5 seconds) Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: our findings suggest that use of aromatherapy with lavender can reduce the pain of blood sampling in term neonates and can be considered as an effective intervention in neonate pain management, although more research is recommended. Study funding source: this project was funded by Mashhad University of Medical Sciences, Mashhad, Iran (grant numbers 88474) Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	High risk	A high‐risk method of sequence generation was described.
2) Selection bias (allocation concealment)	High risk	Due to sequence generation method that was described, researchers would be aware of group assignment and allocation concealment would therefore not be maintained.
3a) Performance bias (blinding of personnel)	High risk	Blinding not possible due to nature of intervention (smell).
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcomes assessor was stated to be blinded.
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data observed.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	High risk	Only one person coded the outcome so no reliability ratings are available.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Ren 2019.

Study characteristics
Methods	Study design: between‐group Total study duration: 9 months
Participants	Total number: 396 Setting: hospital Diagnostic criteria: preterm infants Gestational age: Mean: 30.6 (white noise), 31.3 (glucose), 31.0 (white noise + glucose), 31.2 (control), overall mean = 31.0 SD: 1.3 (white noise), 1.2 (glucose), 1.4 (white noise + glucose), 1.3 (control), overall SD = 1.3   Sex: Males: 47 (white noise), 46 (glucose), 54 (white noise + glucose), 48 (control), total = 195 Females: 49 (white noise), 52 (glucose), 48 (white noise + glucose), 52 (control), total = 201 Country: China
Interventions	Total number of intervention groups: 3 Control group description: the control group was given conventional postural comfort Total number of participants randomised to each of the groups: white noise = 96, glucose = 98, white noise + glucose = 102, control = 100 Specific intervention: white noise, glucose, white noise + glucose Intervention details: White noise: the white noise group was given white noise for intervention. Use the album "Colic Baby: White Noise for Babies" released by Orchard Enterprises as the source of white noise, and use music MP3 external sound player to play in loop. After disinfection, the player is placed in a fixed position 0.5 m away from the premature infant on the desktop of the premature infant examination room. It starts to play 1 minute before the examination and until 5 minutes after the examination. The white noise volume is controlled at 55 decibels with a decibel meter. Glucose: The glucose group was given an oral 25% glucose solution for intervention. A nurse uses a disposable sterile syringe to draw 1 mL of the prepared 25% glucose solution. Before the ROP screening operation, use the syringe nipple to slowly inject the 25% glucose solution into the mouth of the premature infant to suck it. White noise + glucose: the white noise combined with glucose group combines the above intervention methods of the white noise group and the glucose group, that is, the premature infant is slowly injected with 25% glucose solution into the oral cavity while playing the white noise. Integrity of intervention: white noise = good, glucose = good, white noise + glucose = good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: unclear due to language barriers Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: unclear due to language barriers Study funding source: not reported Author COI statement: unclear due to language barriers
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	Unclear risk	It is not clear whether the personnel were blinded to the intervention.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were blinded to study condition.
4) Attrition bias (incomplete outcome data)	Unclear risk	Attrition/exclusions not described so unable to determine whether there was incomplete outcome data.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	Unclear due to language barriers.

Roshanray 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: 2 months
Participants	Total number: 135 Setting: hospital Diagnostic criteria: full‐term healthy neonates Gestational age: Mean: 39.3 (mother's hug), 39.5 (massage), 39.1 (control), overall mean = 39.3 SD: 1.1 (mother's hug), 1.2 (massage), 0.9 (control), overall SD = 1.1   Sex: Males: 21 (mother's hug), 24 (massage), 21 (control), total = 66 Females: 24 (mother's hug), 21 (massage), 24 (control), total = 69 Country: Iran
Interventions	Total number of intervention groups: 2 Control group description: the routine of the health centre (laying on the bed during blooding sampling) was performed. In the routine care mothers are allowed to be present at bedside if they want and they are allowed to keep their baby hands while their babies are on bed during blood sampling. However, pharmacological analgesia is not used routinely. Diapers were changed to reduce stimulating the neonate. Also, all neonates were fed at least 30 minutes before blood sampling. Total number of participants randomised to each of the groups: mother's hug = 45, massage = 45, control = 45   Specific intervention: mother's hug, massage Intervention details: mother's hug: in the mother’s hug, infant was in breastfeeding position, the neonates were placed in their mother’s arms 2 minutes before, during and at least 15 s after blood sampling, with the difference that they were not breast fed. All neonates were awake at the time of applying this method. Their mothers sat on the rocking chairs, and they were allowed to cuddle their baby in case of crying during the blood sampling. Massage: in the massage group, before blood sampling, the researcher gently massaged the baby’s foot from the fingertips to the middle of the leg. The massage was done with the technique of effleurage with the thumb of one hand for 2 minutes to move the contents of the lymph vessels and facilitate blood sampling, and also to move tissue fluids. The effleurage was done in the direction of the venous and lymphatic flow toward the heart. The massage was soft, rhythmic, uniform and constant, and always from the distal point to proximal areas. Integrity of intervention: mother's hug = good, massage = good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: before, immediately, and 5 minutes after blood sampling Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: the placement of the baby in the mother’s hug during painful procedures is recommended due to the reduction of pain, the improvement of physiological symptoms, and the promotion of neonatal health. Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Blinding not mentioned but unlikely due to nature of interventions.
3b) Detection bias (blinding of outcome assessors)	High risk	Blinding not mentioned but unlikely given it was done by a nurse working in the NICU and due to the nature of the intervention.
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data observed.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Sadathosseini 2013.

Study characteristics
Methods	Study design: between‐group Total study duration: 8 weeks to put together sample size
Participants	Total number: 135 Setting: Bahrami Children Hospital Diagnostic criteria: healthy Age: Mean: 5.0 days (familiar scent), 5.0 days (unfamiliar scent), 5.1 (control), overall mean = 5.1 SD: 1.2 (familiar scent), 1.1 (unfamiliar scent), 5.1 (control), overall SD = 1.2   Sex: Males: 23 (familiar), 21 (unfamiliar), 20 (control), total = 64 Females: 22 (familiar), 24 (unfamiliar), 25 (control), total = 71 Country: Iran
Interventions	Total number of intervention groups: 2 Control group description: no scent was given to the control group   Total number of participants randomised to each of the groups: intervention #1 = 45; intervention #2 = 45; control = 45 Specific intervention #1: familiar scent Intervention details #1: familiarisation was performed the night before blood sampling and neonates were exposed to the vanillin scent during the painful procedure. The vanillin solution on the pad was placed near the neonate's head. The next morning, the scented gauze was removed for an average familiarisation of 8.65 hours. After 2 minutes, a sterile gauze pad perfumed with 10 drops of the vanillin solution was held 1 cm from the neonate's nose without contact.   Specific intervention #2: unfamiliar scent Intervention details #2: after 2 minutes, a sterile gauze pad perfumed with 10 drops of the vanillin solution was held 1 cm from the neonate's nose without contact Integrity of intervention: good
Outcomes	Pain outcomes, time points when measured, and scale limits: Cry duration Time points: infant cry was recorded from onset after needle insertion until a cessation for 5 seconds, up to a maximum of 1 minute after removing the needle Range of possible scores: 0 to 60 seconds Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: stimulating neonates with a familiar scent reduced cry duration compared to other groups Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	“Sampling was done only for one group during each week, with groups being randomly selected… On the first day, the name of each group was written on a separate piece of paper, which was enclosed in an envelope, and an independent nurse who worked in the clinical context and was unaware of purpose of the study picked up one envelope randomly as the first group. Then selection was performed the same way for the second group, and the remaining envelope was considered the third group. During each week, eligible cases were recognized and selected by convenience sampling only for one group, and groups were replaced respectively (e.g. first week, unfamiliar scent group; second week, familiar…” Page e197, last paragraph of second column into page e198.
2) Selection bias (allocation concealment)	Low risk	See above.
3a) Performance bias (blinding of personnel)	Low risk	“… the researcher who performed familiarization and olfactory stimulation was the only person who had knowledge of group assignments”. Page e199, 2nd paragraph of 1st column.
3b) Detection bias (blinding of outcome assessors)	Low risk	“… the researcher who performed familiarization and olfactory stimulation was the only person who had knowledge of group assignments”. Page e199, 2nd paragraph of 1st column.
4) Attrition bias (incomplete outcome data)	Unclear risk	It was unclear how missing participants or data were accounted for.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Unclear risk	“Also neonates were excluded if they were not calm before data collection".
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (0/7), unclear risk (2/7).

Secil 2014.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 60 Setting: hospital Diagnostic criteria: healthy Age: Intervention: 8.8 ± 5.7 months Control: 5.7 ± 4.8 months   Sex:  Males: 15 (vibration anaesthesia), 16 (control) = 31 total Females: 15 (vibration anaesthesia), 14 (control) = 29 total Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: no intervention prior to or during venipuncture Total number of participants randomised to each of the groups: Intervention: 30 Control: 30 Specific intervention: vibration anaesthesia Intervention details: a vibration anaesthesia device was placed against the skin, within 5 cm to 10 cm of the venipuncture site, while pressure was gently applied. The device was turned on vibrate for 2 to 3 seconds before and during the procedure. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured FLACC Time points: before, during, and after the procedure Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: a vibration anaesthesia device did not reduce pain scores in infants during and after venipuncture procedure Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	The authors describe a lottery system, which randomly generated the sequence.
2) Selection bias (allocation concealment)	Unclear risk	While the envelopes containing the allocation sequence were opaque, there is no mention of them being sequentially numbered and sealed. More information is required.
3a) Performance bias (blinding of personnel)	High risk	There is no mention of blinding of study personnel. Additionally, it would be very difficult to blind study personnel to study conditions given nature of intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	No mention of blinding of outcome assessors and, given that it was in the moment, would not have been possible.
4) Attrition bias (incomplete outcome data)	Low risk	Researchers recruited/enrolled 60 infants and reported data on 60 infants. No missing outcome data.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall rating is high.

Shaw 1982.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 20 Setting: clinic Diagnostic criteria: healthy children Age: Intervention: M = 62.3 months Control: M = 60.3 months Sex: Intervention: 5 males, 5 females Control: 6 males, 4 females Country: United States
Interventions	Total number of intervention groups: 1 Control group description: the child received routine immunisation as part of a regular visit to a paediatrician's office Total number of participants randomised to each of the groups: Mother‐present group: 10 Control group: 10 Specific intervention: mother‐present Intervention details: mother present: the child was placed in a supine position and, with the parent (in the mother‐present group) or one of the behaviour coders (in the mother‐absent group) holding the child’s hands, was given the DPT injection in the thigh by the nurse. Children in the mother‐present group were immediately picked up by their mothers. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured Modified Frankl Scale/crying duration Time points: the tally sheets used by the behaviour coders divided each part of the examination time into 20‐second intervals. The coders recorded, for each interval, the presence of each of the following behaviours. Range of possible scores: N/A Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: during the immunisation, for those in the mother‐present condition behaviour was rated as significantly more negative than the mother absent group Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	No report of randomisation.
2) Selection bias (allocation concealment)	Unclear risk	No report of allocation bias.
3a) Performance bias (blinding of personnel)	High risk	No way of blinding this if the mother is absent or present.
3b) Detection bias (blinding of outcome assessors)	High risk	No way of blinding this intervention.
4) Attrition bias (incomplete outcome data)	High risk	10 additional participants were recruited and then never reported on.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score on Cochrane standard: low risk (1/7), high risk (3/7), unclear risk (3/7).

Shu 2014.

Study characteristics
Methods	Study design: between participants Total study duration: July 2009 to December 2010
Participants	Total number: 75  Setting: district hospital in Taiwan Diagnostic criteria: healthy neonates (Apgar scores > 7) Age: Minimum: N/A Maximum: N/A Mean: 266.8 days gestation (control), 269.2 days gestation (swaddling), 269.6 days gestation (heel warming), overall mean = 238.5 SD:15.9 (control), 10.9 (swaddling), 10.4 (heel warming), overall SD = 12.4   Sex: Males: 12 (control), 17 (swaddling), 11 (heel warming), total = 40 Females: 13 (control), 8 (swaddling), 14 (heel warming), total = 35 Country: Taiwan
Interventions	Total number of intervention groups: 2 Control group description: routine care: placed in a supine position while lying inside a crib and left without intervention for 30 minutes before performing heel stick   Total number of participants randomised to each of the groups: intervention #1 = 25; intervention #2 = 25; control = 25 Specific intervention #1: swaddling Intervention details #1: neonates were placed face up on a 90 cm blanket. The arms of the neonates were placed close to their torso with both hands clasped. The upper and lower ends of the wrap were kept open, and the upper rim was aligned with the shoulder of the neonate. The horizontal ends of the wrap were folded in opposite directions to cover the torso. An appropriate space was retained at the foot of the wrap before folding it upward to the front, completing the swaddle.   Specific intervention #2: heel warming (treatment arm included in qualitative analysis only) Intervention details #2: superficial heat between 40 to 45 degrees C increases the blood flow. The effects of heat on skin temperature increased rapidly during the first 4 minutes. We put water at 40 degrees C in a thermal bag and applied the bag against the puncture point for 5 minutes. Integrity of intervention: good
Outcomes	Pain outcomes, time points when measured, and scale limits: NIPS Time points: measured immediately after Range of possible scores: 0 to 7, with high scores indicating higher pain reactivity Cry duration Time points: started measuring from the first cry immediately after heel stick to the moment the crying stopped Range of possible scores: not reported Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: swaddling and heel warming reduce the pain response of neonates during the heel stick procedure ‐ heel warming resulted in lower pain response than swaddling during pain recovery Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Did not specify.
2) Selection bias (allocation concealment)	Unclear risk	Did not specify.
3a) Performance bias (blinding of personnel)	Unclear risk	Did not specify.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	Did not specify.
4) Attrition bias (incomplete outcome data)	Low risk	Depicted in Figure 1 on page 3109.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (3/7), high risk (0/7), unclear risk (4/7).

Sizun 2002.

Study characteristics
Methods	Study design: cross‐over Total study duration: not reported
Participants	Total number: 19 Setting: hospital NICU Diagnostic criteria: preterm neonates, younger than 32 weeks Age: Minimum: unknown Maximum: unknown Mean: 29.0 weeks SD: 1.8 weeks Sex: Males: 10 Females: 9 Country: France
Interventions	Total number of intervention groups: 1 Control group description: no protection from light, supine, no supportive bedding, no individualised attention Total number of participants randomised to each of the groups: 19 Specific intervention #1: developmental care Intervention details: decreased light, decreased noise, lateral posture, feet contacting supportive bedding, opportunities for grasping, offered motor support by hard swaddling, grasping, support for state transition Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured: PIPP Time points: immediate Upper limit: 21 Lower limit: 0 Higher score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: developmental care works on reactivity Study funding source: supported by PHRC 1998 grant Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Unclear risk	Randomisation is mentioned but there is not an adequate description of the methods used.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Unclear risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	High risk	There is no documented evidence or insufficient evidence reported of how attrition was dealt with.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (1/7), high risk (1/7), unclear risk (5/7).

Stevens 1999.

Study characteristics
Methods	Study design: cross‐over Total study duration: 15 months
Participants	Total number: 122 Setting: NICU Diagnostic criteria: preterm infants Age: Minimum: 27 weeks Maximum: 31 weeks Mean: 28.6 weeks SD: 0.2 weeks Sex: Males: 70  Females: 52  Country: Canada Co‐morbidity: none Socio‐demographics: not reported Ethnicity: 78 white, 12 black, 17 Hispanic, 15 Asian
Interventions	Total number of intervention groups: 2 Control group description: in a Snuggle‐Up device for 30 minutes prior to heel lance procedure. Infant was side‐lying or in the supine position in the Snuggle‐Up. Total number of participants randomised to each of the groups: 122 Specific intervention #1: prone positioning Intervention details: infants were positioned prone in Snuggle‐Up, with knees flexed, arms in, arms close to midline and left foot free Integrity of intervention: good Specific Intervention #2: pacifier with water Intervention details: infants were given a pacifier dipped in water and positioned in Snuggle‐Up for 7 minutes pre lance Integrity of intervention: good
Outcomes	Pain outcomes, time points when measured, and scale limits: PIPP Time points: immediately (stick 15 seconds), 5 minutes after Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: pacifier with water was the most effective method in reducing pain Study funding source: financial support is acknowledged from the National Institutes of Health, National Institute of Nursing Research (grant NR 03916 R01) and a National Institutes of Health Pediatric Clinical Research Center grant at the University of California San Francisco and Children's Hospital, Oakland research sites (M01 ‐ RR012771) Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Unclear risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (3/7), high risk (1/7), unclear risk (3/7).

Sundaram 2013.

Study characteristics
Methods	Study design: cross‐over Total study duration: less than 1 month
Participants	Total number: 20 Setting: NICU Diagnostic criteria: Preterm newborns of both sexes between 28 and 36 weeks of GA who were breathing unassisted, not receiving paralytic, analgesic, or sedative medications within 48 hours, without major congenital anomalies, not suffering from grade III or IV intraventricular haemorrhage (IVH) or subsequent periventricular leukomalacia (PVH), had not undergone any surgery from a level II NICU of a tertiary care hospital through purposive sampling, and from whom parental consent (approved by the institution) was obtained after explaining the purpose and procedure of the study   Age: Minimum: 28 weeks Maximum: > 36 weeks gestation Mean: 34.1 weeks gestation; 7.8 days SD: 2.3 weeks gestation; 7.7 days Sex: Males: 11 Females: 9 Country: India
Interventions	Total number of intervention groups: 1 Control group description: standard care Total number of participants randomised to each of the groups: 20 Specific intervention: facilitated tucking Intervention details: described as holding the infant by placing a hand on his or her hands and feet, positioning them in a flexed midline position while in either a side‐lying, supine, or prone position Integrity of intervention: good
Outcomes	Pain outcomes, time points when measured, and scale limits: PIPP Time points: 30, 60, 90, 120 seconds post heel stick Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: facilitated tucking is an effective non‐pharmacological pain management in preterm newborns during routine neonatal intensive care
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported that used an independent person not involved in enrolment or allocation of participants.
2) Selection bias (allocation concealment)	Low risk	There is adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Unclear risk	Not specified.
3b) Detection bias (blinding of outcome assessors)	Low risk	Coders were blinded.
4) Attrition bias (incomplete outcome data)	Unclear risk	Participants that had missing data were simply eliminated from the analyses. It was unclear how missing participants or data were accounted for.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Unclear risk	As per Sterne 2019, total score based on Cochrane standard: low risk (4/7), high risk (0/7), unclear risk (3/7).

Taddio 2014.

Study characteristics
Methods	Study design: between‐group Total study duration: 16 months
Participants	Total number: 121 Setting: private paediatric clinic Diagnostic criteria: full‐term health infants Postnatal age: Mean: 165 days (tactile stimulation), 162 days (control) SD: 121 (tactile stimulation), 122 (control) Sex: Males: 68 Females: 53 Country: Canada
Interventions	Total number of intervention groups: 1 Control group description: both groups received 2 mL of sucrose 24% solution and received intramuscular injections administered rapidly without prior aspiration   Total number of participants randomised to each of the groups: Tactile stimulation: 62 Control: 59   Specific intervention: tactile stimulation Description of intervention: parents premedicated infants with 2 mL of sucrose 24% solution. Injections were performed in the anterolateral aspect of the right or left thigh using a 25‐gauge, 22 mm needle. Intramuscular injections were administered rapidly without prior aspiration. Additionally, the paediatrician additionally rubbed the skin adjacent to the injection site with moderate intensity using the non‐injection thumb for 5 to 10 seconds preceding and during vaccination and 5 to 10 seconds after vaccination.   Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured MBPS Time points: first 30 seconds post‐procedure Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse reactions: in all, 28 (23%) infants were unsettled during administration of sucrose, and in 4 of them, administration of sucrose was aborted because of infant distress. In 2 others, the dose was repeated without incident after infants were allowed to settle. Three infants spit up during or after sucrose administration.
Notes	Key conclusions of study authors: tactile stimulation cannot be recommended as a strategy to reduce vaccination pain in infants because of insufficient evidence of a benefit Study funding source: Pfizer, Ferndale and Natus, summer undergraduate research programme at the Leslie Dan Faculty of Pharmacy, University of Toronto, summer undergraduate research programme at The Hospital for Sick Children (SickKids) Author COI statement: the author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: AT was involved in conception and design; analysis and interpretation of data; drafting of manuscript; critical revision; funding; administrative, technical, and material support; and supervision. AT had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. TH, M‐EH, and VS were involved in conception and design, interpretation of data, and critical revision of the manuscript. CV, AJ, ST, and AFI were involved in aspects of design, acquisition or analysis of data, interpretation of data, and critical revision of the manuscript. Sucrose 24% solution supplies for the trial were provided by Equinox. The other authors have no conflicts of interest relevant to this article to disclose.
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	Paediatricians were not blinded as they were required to deliver the intervention. However, protocol for delivering immunisation was standardised and so lack of blinding is not expected to have affected delivery of intervention.
3b) Detection bias (blinding of outcome assessors)	Low risk	MBPS observer was blinded to study condition.
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Low risk	As per Sterne 2019, if all domains received a rating of low risk then the overall assessment is low risk.

Taddio 2015.

Study characteristics
Methods	Study design: between‐participant Total study duration: not reported
Participants	Total number: 160 Setting: St. Michael’s Hospital, Toronto, ON Diagnostic criteria: healthy Age: Mean: 75 days (parent education), 81 days (control), overall mean = 78 SD: 28 (parent education), 29 (control), overall SD = 28.5 Sex: Males: 40 (parent education), 48 (control), total = 88 Females: 40 (parent education), 32 (control), total = 72 Country: Canada
Interventions	Total number of intervention groups: 1 Control group description: parents in the control group received general information on the diseases infant immunisations protect against adopted from the Toronto Public Health website   Total number of participants randomised to each of the groups: Intervention #1: parent education Control: 80 Intervention description: parents in the intervention group received information on how they can comfort their infants during immunisations in the form of written and video information Integrity of intervention: good
Outcomes	Pain outcomes, time points when measured, and scale limits: MBPS Time points: 15 seconds post‐needle Range of possible scores: 0 to 10 (higher numbers, greater pain) Cry time Time points: 120 seconds post‐needle Range of possible scores: 0 to 120 seconds Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: parent education can reduce pain regulation in children during immunisations and parents should be taught about non‐pharmacological strategies Study funding source: this study was funded by miscellaneous funds held by A. Taddio. C. Parikh was funded, in part, by a Trainee Award from a Canadian Institutes of Health Research (CIHR) Pain in Child Health Training Consortium. A. Taddio received research funding from Pfizer and study supplies from Ferndale and Natus for a separate clinical trial related to this topic. R. Pillai Riddell is supported by a New Investigator Award from CIHR. E. Yoon is supported by the Ontario Ministry of Health and Long‐Term Care. Author COI statement: this study was funded by miscellaneous funds held by A. Taddio. C. Parikh was funded, in part, by a Trainee Award from a Canadian Institutes of Health Research (CIHR) Pain in Child Health Training Consortium. A. Taddio received research funding from Pfizer and study supplies from Ferndale and Natus for a separate clinical trial related to this topic. R. Pillai Riddell is supported by a New Investigator Award from CIHR. E. Yoon is supported by the Ontario Ministry of Health and Long‐Term Care. The other authors declare no conflicts of interest relevant to this article.
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	Using a computer random number generator, a research assistant (RA) uninvolved with the study retrieved a randomisation code for the 4 participating physician practices from the study statistician – 1/2 in control and 1/2 in intervention group.
2) Selection bias (allocation concealment)	Low risk	Did not need this as it was a cluster trial.
3a) Performance bias (blinding of personnel)	High risk	“The members of the clinical staff and the candidate involved with recruitment, education, and follow‐up could not be blinded to treatment allocation due to the nature of the intervention”; “Parents were blinded to their group allocation.”
3b) Detection bias (blinding of outcome assessors)	Low risk	“The trained RAs that scored strategy utilization, pain, and interactions were blinded to the purpose of the study and the treatment allocation. Study personnel collecting data and clinicians were instructed to ensure that any indication of group allocation is not captured during the video recording so as to maintain concealment of treatment allocation.”
4) Attrition bias (incomplete outcome data)	Low risk	Depicted in Figure 5.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (5/7), high risk (1/7), unclear risk (1/7).

Taddio 2017.

Study characteristics
Methods	Study design: between‐group Total study duration: 49 months
Participants	*Numbers reported are only those included in review; due to longitudinal nature of study, only the 6‐month old infants were included as this age group was closest to other infant age groups in the analysis Total number: 156 Setting: clinic Diagnostic criteria: healthy full‐term infants Postnatal age: Mean: 191.0 days (parent education), 191.8 days (control) SD: 15.8 (parent education), 11.2 (control)   Sex: Males: 51 (parent education), 46 (control), total = 97 Females: 38 (parent education), 42 (control), total = 80 Country: Canada
Interventions	Total number of intervention groups: 1 Control group description: at 2, 4, 6, and 12 months, parents received sham educational video (general information about immunisation pain management), sucrose placebo (2 mL water one to 2 minutes pre‐needle), and lidocaine placebo (1 g topical cream applied to injection site(s) for 20 minutes pre‐needle) Total number of participants randomised to each of the groups: Intervention: 75 Control: 81 Specific intervention: parent education Intervention description: at 2, 4, 6, and 12 months, parents received active educational video (parent‐directed video education about infant soothing), sucrose placebo (2 mL water one to 2 minutes pre‐needle), and lidocaine placebo (1 g topical cream applied to injection site(s) for 20 minutes pre‐needle) Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured MBPS Time points: first 30 seconds post‐procedure, 2 minutes post‐procedure Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse reactions: transient skin reactions to cream (pallor/redness) were common, ranging from 32.1% to 71.8% (pallor) and 58.5% to 80.2% (redness) across all time points
Notes	Key conclusions of study authors: no difference in pain scores between the 2 groups Study funding source: Pfizer Author COI statement: Anna Taddio declares a research grant from Pfizer, and study supplies from Natus and Ferndale. The other authors declare no conflicts of interest.
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Low risk	An adequate method is reported that removes potential biases of investigators.
3a) Performance bias (blinding of personnel)	Low risk	Personnel were blinded to study condition.
3b) Detection bias (blinding of outcome assessors)	Low risk	All participants received an active or placebo form of all the interventions (and were blinded to whether they had received active/placebo forms), and thus all outcome assessors (parents, physicians, researchers, MBPS coders) would be unaware of treatment condition.
4) Attrition bias (incomplete outcome data)	Low risk	Despite dropout rates being slightly uneven between groups (more in video‐sucrose‐lidocaine condition), retention rates ranged from 82% to 85% overall, suggesting good retention. Additionally, reasons for withdrawal are documented.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Low risk	As per Sterne 2019, if all domains received a rating of low risk then the overall assessment is low risk.

Taddio 2017a.

Study characteristics
Methods	Study design: between‐group Total study duration: 49 months
Participants	Total number: 352 Setting: clinic Diagnostic criteria: healthy full‐terms Age: 2 months, 4 months, 6 months, 12 months, 15 months Sex: 192 males, 160 females Country: Canada
Interventions	Total number of intervention groups: 3 Control group description: at 2, 4, 6, and 12 months, received sham educational video (general information about immunisation pain management), sucrose placebo (2 mL water on tongue 2 minutes pre‐needle), and lidocaine placebo (1 g cream applied to injection site(s) for 20 to 30 minutes pre‐needle). At 15 months, received all 3 active interventions (educational video, sucrose, lidocaine). Also received pain‐relieving techniques that are part of routine care. Total number of participants randomised to each of the groups: Video = 89, video + sucrose = 88, video + sucrose + lidocaine = 87 Specific intervention: video, video + sucrose, video + sucrose + lidocaine Description of intervention:  Video: at 2, 4, 6, and 12 months received active educational video (techniques related to optimal soothing), sucrose placebo (2 mL water on tongue 2 mins pre‐needle), and lidocaine placebo (1 g cream applied to injection site(s) for 20 to 30 minutes pre‐needle). At 15 months, received all 3 active interventions (educational video, sucrose, lidocaine). Also received pain‐relieving techniques that are part of routine care (administration of the most painful vaccine last, intramuscular injection without aspiration, and holding infants). Video + sucrose: at 2, 4, 6, and 12 months, received active educational video (techniques related to optimal soothing), sucrose (2 mL oral sucrose 24% solution on tongue 2 minutes pre‐needle), and lidocaine placebo (1 g cream applied to injection site(s) for 20 to 30 minutes pre‐needle). At 15 months, received all 3 active interventions (educational video, sucrose, lidocaine). Also received pain‐relieving techniques that are part of routine care (administration of the most painful vaccine last, intramuscular injection without aspiration, and holding infants). Video + sucrose + lidocaine: at 2, 4, 6, and 12 months, received active educational video (techniques related to optimal soothing), sucrose (2 mL oral sucrose 24% solution on tongue 2 minutes pre‐needle), and lidocaine (1 g topical liposomal lidocaine 4% cream applied to injection site(s) for 20 to 30 minutes pre‐needle). At 15 months, received all 3 active interventions (educational video, sucrose, lidocaine). Also received pain‐relieving techniques that are part of routine care (administration of the most painful vaccine last, intramuscular injection without aspiration, and holding infants). Integrity of intervention: video = good, video + sucrose = good, video + sucrose + lidocaine = good
Outcomes	Pain outcomes, scale limits, and time points when measured MBPS Time points: pre‐procedure (baseline), after vaccine injection(s) (needle), and 1 minute after the last vaccine injection (recovery) Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: there was no evidence of a long‐term benefit of consistent use of pain interventions in the first year of life on future infant pain responsivity at 15‐month vaccinations Study funding source: external funding was acquired by A.T. from Pfizer via the Investigator Initiated Research program. Study supplies (placebo and active) were supplied by Ferndale, Natus, and Aboutkidshealth (The Hospital for Sick Children). Author COI statement: A.T. declares a research grant from Pfizer, and study supplies from Natus, Ferndale, and Aboutkidshealth. The remaining authors have no conflicts of interest to declare.
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Low risk	A convincing method for allocation concealment is reported.
3a) Performance bias (blinding of personnel)	Low risk	Personnel were blinded to study condition.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were blinded to study condition.
4) Attrition bias (incomplete outcome data)	Low risk	The number of enrolled participants in each group and the number reported in outcome tables match.
5) Selective reporting (reporting bias)	Low risk	We accessed the study protocol and all pre‐specified outcomes were reported in the study.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	Low risk	As per Sterne 2019, if all domains fall under low risk, the overall assessment is low risk.

Taplak 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: 12 months
Participants	Total number: 80 Setting: hospital Diagnostic criteria: preterm infants requiring mechanical ventilation Gestational age: 43 were between 26 to 31.6 weeks; 37 were between 32 to 35.6 weeks Sex: Males: 10 (breast milk smell), 10 (white noise), 8 (facilitated tucking), 9 (control) total = 37 Females: 10 (breast milk smell), 10 (white noise), 12 (facilitated tucking), 11 (control), total = 43 Country: Turkey
Interventions	Total number of intervention groups: 3 Control group description: infants who underwent the endotracheal suctioning procedure. The routine suctioning procedure is as follows: preterm infant is on the right side and aspirate using an open suctioning system. The patient is gently disconnected from the ventilator. The suction catheter is withdrawn gently applying continuous suction pressure by placing the thumb over the suction control port. The saline or hyperoxygenation administration before suctioning is performed not routinely but as needed by the infant. The suction negative pressure is applied not to exceed 100 mm Hg. Total number of participants randomised to each of the groups: breast milk smell = 20, white noise = 20, facilitated tucking = 20, control = 20 Specific intervention: breast milk smell, white noise, and facilitated tucking Intervention details: Breast milk smell: for the inhalation method, 1 mL of breast milk was dropped on apet/filter paper. The smell sample was placed 10 cm away from the infant's nose. The breast milk from the preterm infants' mothers was used as the sample, and for this purpose, the infants' mothers were contacted, and breast milk (in the form of transitional milk) was obtained to be used on the procedure day. White noise: in the selection of music to be played for the preterm infants, expert opinion was obtained. The white noise composed of the mother's womb environment and real heartbeat sounds and relaxing music. The infant listened to the piece through 2 mini loud speakers of an MP3 player that was placed to the right and left of the infant's feet. A decibel meter was placed near the head of the preterm infant to measure the sound level of the music, and the maximum volume was set at 45 dB. Facilitated tucking: the preterm infants were held in the facilitated tucking position by a nurse working in the service. During the procedure, to prevent heat loss, the lid of the incubator was not opened; instead, the nurse accessed the infant through the access port doors of the incubator. Facilitated tucking was performed in such a way that the preterm infant was laid on their right side with the arms and legs flexed and the knees and elbow tucked to midline. If an infant's head is tilted down when he or she is on mechanical ventilator, this causes respiratory problems; therefore, the infant's head was held in the straight position. Integrity of intervention: breast milk smell = satisfactory, white noise = satisfactory, facilitated tucking = satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP‐R Time points: 1, 3, 5 minutes before endotracheal suctioning procedure, during procedure, and 1, 3 and 5 minutes after the endotracheal suctioning procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: white noise and facilitated tucking were effective in relieving pain before procedure and facilitated tucking recovery in preterm infants after the ES procedure Study funding source: the authors received no financial support for the research, authorship, and/or publication of this article Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel were not blinded to group allocation.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were stated to be blinded to group allocation.
4) Attrition bias (incomplete outcome data)	Unclear risk	Unclear reasons for attrition of some participants. It was unclear how missing participants or data were accounted for.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Thakkar 2016.

Study characteristics
Methods	Study design: between‐group Total study duration: 12 months
Participants	Total number: 180 Setting: hospital (NICU) Diagnostic criteria: full‐term neonates Gestational age: 68 were 38 weeks gestation 91 were 39 weeks gestation 21 were 40 weeks gestation Sex: 103 males and 77 females Country: India
Interventions	Total number of intervention groups: 3 Total number of participants randomised to each of the groups: sucrose = 45; NNS = 45; sucrose + NNS = 45; control = 45 Description of control group: the neonates were placed naked in a cot under a radiant warmer in a quiet, diffusely lighted room. No other intervention was given. Integrity of intervention: good   Intervention descriptions: Specific intervention #1: sucrose Description of intervention #1: the neonates were placed naked in a cot under a radiant warmer in a quiet, diffusely lighted room. The pulse oximeter probe was attached to the right hand. Received 30% sucrose solution by sterile syringe. Integrity of intervention: good   Specific intervention #2: NNS Description of intervention #1: the neonates were placed naked in a cot under a radiant warmer in a quiet, diffusely lighted room. The pulse oximeter probe was attached to the right hand. Sterile gauze was held gently in neonate’s mouth and the palate tickled to stimulate sucking.  Integrity of intervention: good   Specific intervention #3: sucrose + NNS Description of intervention #3: the neonates were placed naked in a cot under a radiant warmer in a quiet, diffusely lighted room. The pulse oximeter probe was attached to the right hand. Immediately before heel stick, received 30% sucrose solution by sterile syringe and had sterile gauze gently held in mouth and palate tickled to stimulate sucking. All neonates experienced squeezing (< 30 seconds) to collect the sample. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: a total of 5 episodes of adverse events occurred in all the 4 groups. One neonate each in group I, II and III desaturated, whereas 2 neonates desaturated in the no intervention group (group IV). No episode of abnormal heart rate was noted in any of the groups.
Notes	Key conclusions of study authors: sucrose and/or NNS are effective in providing analgesia in full‐term neonates undergoing heel stick procedures, with the combined intervention being more effective compared with any single intervention Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	While authors note that the envelopes were opaque and sealed, there is no mention of whether envelopes were sequentially numbered. There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Study personnel were unable to be blinded to intervention condition due to overt nature of NNS.
3b) Detection bias (blinding of outcome assessors)	High risk	While there was no mention of whether outcome assessors were blinded, it would not have been possible given the overt nature of the NNS.
4) Attrition bias (incomplete outcome data)	Low risk	180 neonates were randomised, and data on 180 neonates were reported. No missing outcome data.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under "high risk" then the overall assessment is high risk.

Ti 2015.

Study characteristics
Methods	Study design: between‐group Total study duration: not mentioned
Participants	Total number: 110 Setting: level III neonatal intensive care Diagnostic criteria: healthy neonates Gestational age (weeks): 27 to 37 Mean: 32.5  SD: 3.1 Sex:  Males: 49.5% Females: 50.5% Country: Taiwan
Interventions	Total number of intervention groups: 5 Control group description: 1 hour before heel stick, regardless of treatment condition, infants were placed in a prone, lateral, or supine position with support. Before, during, and after heel stick, infants in the control condition (routine care) received only gentle touch and verbal comfort if they were crying.   Total number of participants randomised to each of the groups: Non‐nutritive sucking + facilitated tucking: 22 Oral sucrose + facilitated tucking: 21 Non‐nutritive sucking + oral sucrose: 21 Non‐nutritive sucking + oral sucrose + facilitated tucking: 23 Control: 23   Specific intervention #1: non‐nutritive sucking (NNS) + facilitated tucking (FT) Intervention details #1: 1 hour before heel stick, infants were placed in a prone, lateral, or supine position with support. Before, during, and after heel stick, infants received non‐nutritive sucking + facilitated tucking. For the NNS condition, infants received a standard silicone newborn pacifier 1 minute before their foot was touched to start heel stick procedures. For the FT condition, infants were gently held in warm hands, keeping the infant in a flexed, midline position with his/her 4 limbs close to his/her own body and minimum restraint of the infant’s head and body. Integrity of intervention #1: good Specific intervention #2: oral sucrose + facilitated tucking (FT) Intervention details #2: 1 hour before heel stick, infants were placed in a prone, lateral, or supine position with support. Before, during, and after heel stick, infants received oral sucrose + facilitated tucking. For the oral sucrose condition, infants received 0.2 mL to 2.0 mL of 20% sucrose through a syringe 2 minutes before heel stick. The volume of sucrose was fed to the infant based on GA (GA 26 to 28 weeks: 0.2 mL; GA 28.1 to 30 weeks: 0.5 mL; GA 30.1 to 32 weeks: 1 mL; GA 32.1 to 37 weeks: 1.5 mL; GA > 37 weeks: 2.0 mL). For the FT condition, infants were gently held in warm hands, keeping the infant in a flexed, midline position with his/her 4 limbs close to his/her own body and minimum restraint of the infant’s head and body.  Integrity of intervention #2: good   Specific intervention #3: non‐nutritive sucking (NNS) + oral sucrose Intervention details #3: 1 hour before heel stick, infants were placed in a prone, lateral, or supine position with support. Before, during, and after heel stick, infants received non‐nutritive sucking + oral sucrose. For the NNS condition, infants received a standard silicone newborn pacifier 1 minutes before their foot was touched to start heel stick procedures. The volume of sucrose was fed to the infant based on GA (GA 26 to 28 weeks: 0.2 mL; GA 28.1 to 30 weeks: 0.5 mL; GA 30.1 to 32 weeks: 1 mL; GA 32.1 to 37 weeks: 1.5 mL; GA > 37 weeks: 2.0 mL).  Integrity of intervention #3: good Specific intervention #4: non‐nutritive sucking (NNS) + oral sucrose + facilitated tucking (FT) Intervention details #4: 1 hour before heel stick, infants were placed in a prone, lateral, or supine position with support. Before, during, and after heel stick, infants received  non‐nutritive sucking + oral sucrose + facilitated tucking. For the NNS condition, infants received a standard silicone newborn pacifier 1 minute before their foot was touched to start heel stick procedures. The volume of sucrose was fed to the infant based on GA (GA 26 to 28 weeks: 0.2 mL; GA 28.1 to 30 weeks: 0.5 mL; GA 30.1 to 32 weeks: 1 mL; GA 32.1 to 37 weeks: 1.5 mL; GA > 37 weeks: 2.0 mL). For the FT condition, infants were gently held in warm hands, keeping the infant in a flexed, midline position with his/her 4 limbs close to his/her own body and minimum restraint of the infant’s head and body Integrity of intervention #4: good
Outcomes	Pain outcomes, scale limits, and time points when measured Grimace behaviours  Time points: 30 minutes pre‐procedure, during, 30 minutes post‐procedure  Upper limit: not mentioned  Lower limit: not mentioned  High score = more pain Continuous Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors: the combined use of non‐pharmacological interventions (non‐nutritive sucking + oral sucrose + facilitated tucking) more effectively reduced the frequency of infants’ stress related or withdrawal behaviours (grimace and limb and trunk extension or squirming) than routine care across heel stick procedures Study funding source: Ministry of Science and Technology of Taiwan, R.O.C. for grant support (NSC 99‐2314‐B‐016‐039) Author COI statement: there are no conflicts of interest in this study
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel unable to be blinded due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	Low risk	Outcome assessors were blinded to study condition.
4) Attrition bias (incomplete outcome data)	Low risk	The number of enrolled participants in each group and the number reported in outcome tables match.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain is "high", the overall judgement is high.

Vu‐Ngoc 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: 1 month
Participants	Total number: 42 Setting: hospital Diagnostic criteria: both groups are healthy, term newborns Age: Mean: 45.7 h (non‐nutritive sucking), 49.1 h (control), overall mean = 47.3 h SD: 12.8 h (non‐nutritive sucking), 9.0 h (control), overall SD = 11.2 h Sex: Males: 12 (non‐nutritive sucking), 12 (control), total = 24 Females: 10 (non‐nutritive sucking), 8 (control), total = 18 Country: Vietnam
Interventions	Total number of intervention groups: 1 Control group description: newborns in the control group did not receive any form of pain relief Total number of participants randomised to each of the groups: Intervention: 22 Control: 20 Specific intervention: non‐nutritive sucking Intervention details: NNS was performed 120 seconds before the heel prick, during the heel prick, and then for a further 120 seconds after the heel prick. A medical student held the pacifier and applied a gentle pressure to keep it in the infant's mouth during the procedure. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured N‐PASS Time points: 30, 60, 90, and 120 seconds after the heel prick Upper limit: 0 Lower limit: 10 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: NNS is considered a safe and effective pain relief method during the heel prick procedure in term neonates, demonstrated using the N‐PASS. These findings were correlated with the pain evaluation by NFCS and NIPS. Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	Unclear risk	There is not an adequate description of allocation concealment.
3a) Performance bias (blinding of personnel)	High risk	Personnel were stated not to be blinded due to nature of intervention.
3b) Detection bias (blinding of outcome assessors)	High risk	Outcome assessors were not blinded.
4) Attrition bias (incomplete outcome data)	Low risk	Reason for missing outcome data explained and unlikely to be related to true outcome.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Ward‐Larson 2004.

Study characteristics
Methods	Study design: cross‐over Total study duration: not reported
Participants	Total number: 40 Setting: NICU Diagnostic criteria: very preterm infants Age: Minimum: 23 weeks Maximum: 32 weeks Mean: 27.3 weeks SD: 2.4 weeks Sex: Males: 22 Females: 18 Country: USA Co‐morbidity: none Socio‐demographics: none listed Ethnicity: 23 African American, 16 Caucasian, 1 Hispanic
Interventions	Total number of intervention groups: 1 Control group description: infants were put in a Snuggle‐Up (provides containment). No hands were on the infants. Total number of participants randomised to each of the groups: 40 Specific intervention: facilitated tucking Intervention details: infants were put in a Snuggle‐Up and turned to their sides. Their backs were curled gently, legs flexed at a 90 degree angle and their shoulders were brought to midline. They remained like this for suctioning and for 30 seconds after. There was a 2‐ to 4‐hour washout period between procedures. Integrity of intervention: good
Outcomes	Pain outcomes, time points when measured, and scale limits: PIPP Time points: measured for 30 seconds after suctioning Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Funding sources: not mentioned Key conclusions of study authors: facilitated tucking works Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	High risk	Randomisation is mentioned but there is not an adequate description of the methods used.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Unclear risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Unclear risk	Investigator did the coding.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (1/7), high risk (2/7), unclear risk (4/7).

Whipple 2004.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 40 Setting: NICU‐SCN Level III of Hospital in Atlanta, Georgia Diagnostic criteria: low birth weight (< 2500 g), hospitalised in NICU, preterm Age: Minimum: 24.4 weeks Maximum: 35.4 weeks Mean: 31.2 weeks SD: not reported for whole sample Sex: Males: 30 Females: 30 Country: USA
Interventions	Total number of intervention groups: 2 Control group description: no contact, infants were not provided with a pacifier or music listening opportunities at any point during heel stick procedure. Standard care and pain management procedures (such as swaddling, cuddling, and sucrose) were not limited to infants of any of the groups. Total number of participants randomised to each of the groups: 20 in each group Specific intervention #2: pacifier provided, no music Intervention details: sucking intervention began 3 minutes prior to heel stick and continued throughout the procedure until approximately 3 minutes after blood collection Integrity of intervention: satisfactory
Outcomes	TimPain outcomes, scale limits, and time points when measured: Behavioural state (scored based on the "Assessment of Premature Infant's Behavior") Time points: 1‐second intervals and 1 minute pre‐needle, last minute pre‐needle, 1 minute during needle, 1 minute post‐needle, last minute post‐needle Upper limit: 254 Lower limit: 15 High score = high pain Continuous Stress level Time points: 1‐second intervals and 1‐minute pre‐needle, last minute pre‐needle, 1 minute during needle, 1 minute post‐needle, last minute post‐needle Upper limit: 837 Lower limit: 54 High score = high pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: music reinforced NNS, effectively lowers behavioural states and stress levels in low birthweight, preterm infants undergoing heel stick Study funding source: Ohmeda Medical and provision of the Sondrex Pacifier Activated Lullaby (PAL) System Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	High risk	Randomisation is mentioned but there is not an adequate description of the methods used.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Unclear risk	None specified.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	High risk	There is no documented evidence or insufficient evidence reported of how attrition was dealt with.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (1/7), high risk (3/7), unclear risk (3/7).

Yilmaz 2010.

Study characteristics
Methods	Study design: between‐group Total study duration: 12 months
Participants	Total number: 60 Setting: hospital Diagnostic criteria: low birth weight (< 2500 g), Apgar score greater than or equal to 8, fed at least 30 minutes before procedure, no signs of disease or congenital anomalies, C‐section delivery, blood sampling at the first attempt, mothers holding baby in their arms during the procedure Age: 37 to 42 weeks (gestational age) Minimum: 37 weeks GA Maximum: 42 weeks GA Mean: 3.3 days SD: 1.2 Sex: Males: 61 Females: 59 (49.2%) Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: baby in mother's lap, no interventions Total number of participants randomised to each of the groups: 30 Specific intervention #1: pacifier Intervention details: babies were given a pacifier Integrity of intervention: satisfactory
Outcomes	Pain outcomes, scale limits, and time points when measured: NIPS Time points: coded for 2 minutes before, 5 minutes during, and 3 minutes after injections Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: pacifying shortened the length of crying time and the rate of behavioural responses to pain compared to the control group Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	High risk	Randomisation is mentioned but there is not an adequate description of the methods used.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of attempts to deal with potential allocation bias.
3a) Performance bias (blinding of personnel)	Unclear risk	Efforts to reduce measurement bias are not reported or are insufficient, e.g. outcomes collected by therapist.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	None specified.
4) Attrition bias (incomplete outcome data)	Low risk	There are no attrition concerns after comparing number of participants randomised and number of participants included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, total score based on Cochrane standard: low risk (2/7), high risk (2/7), unclear risk (3/7).

Yilmaz 2020.

Study characteristics
Methods	Study design: between‐group Total study duration: not reported
Participants	Total number: 120 Setting: hospital Diagnostic criteria: healthy, full‐term Prenatal age (weeks): Mean: 38.9 (swaddling), 39.0 (swaddling + holding), 39.3 (swaddling + holding + breastfeeding), 38.9 (control), overall mean = 39.0 SD: 1.0 (swaddling), 1.2 (swaddling + holding), 1.0 (swaddling + holding + breastfeeding), 1.2 (control), overall SD = 1.1   Sex: Males: 80 Females: 80 Country: Turkey
Interventions	Total number of intervention groups: 3 Control group description: newborns were in a routine position. No intervention was performed for the newborns during the procedure. Total number of participants randomised to each of the groups: swaddling = 40, swaddling + holding = 40, control = 40 Specific intervention #1: swaddling Intervention details: in compliance with the newborn anatomic posture, the legs were wrapped in the flexion and abduction position. The arms of the newborns were placed close to their body so as not to restrict their movement. Swaddling was carried out 1 minute before the heel stick procedure and continued 2 minutes after the procedure. Integrity of intervention #1: swaddling = good Specific intervention #2: swaddling + holding Intervention details: the mothers in this group sit down on a comfortable chair and following the safe swaddling of the newborns, the baby is placed in the mother's arms and the heel lance process is applied. The newborn is kept in this same position for 2 more minutes following the process. Integrity of intervention #2: swaddling + holding = good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: during procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial did not report on adverse events
Notes	Key conclusions of study authors: all 3 methods are effective in reducing the pain felt during heel lancing in the newborn. However, swaddling, holding, and breastfeeding is more effective than the other methods. Study funding source: not reported Author COI statement: no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	Low risk	A convincing method for generating a random allocation sequence is reported.
2) Selection bias (allocation concealment)	High risk	There is not an adequate description of allocation concealment, as authors referred to randomisation of only 3 groups even though 4 groups present.
3a) Performance bias (blinding of personnel)	High risk	Due to the differences in the intervention types (e.g. swaddled versus not swaddled), it was not possible to blind study personnel to intervention status.
3b) Detection bias (blinding of outcome assessors)	Unclear risk	Included comment on outcome assessors blinding but more detail needed on how blinding was done as the non‐pharmacological interventions given would interfere with the ability to blind.
4) Attrition bias (incomplete outcome data)	Low risk	Provided flow chart showing that all participants that were randomised were included in analyses.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, given that the rating of at least one domain was high risk, the overall rating was high.

Yu 2017.

Study characteristics
Methods	Study design: cross‐over Total study duration: not reported
Participants	Total number: 17 Setting: hospital Diagnostic criteria: healthy full‐term neonates Gestational age: Mean: 38.7 SD: 0.8 Postnatal age: Mean: 4.8 SD: 1.4   Sex: Males: 52.9% Females: 47.1% Country: Hong Kong
Interventions	Total number of intervention groups: 1 Total number of participants randomised to each of the groups: 17 Description of control: no pain control Intervention descriptions: Specific intervention #1: manual pressure Description of intervention #1: the researcher pressed on the neonate’s heel for 10 seconds immediately before the heel prick, using the first digital portion of the thumb. The area under digital pressure was approximately 2.5 x 1.5 = 3.75 cm², measured with a ruler. The researcher then pressed on a baby scale 10 times and recorded the force of every press to the nearest gram. The range and average pressure applied was converted from kg/cm² into kilopascal (kPa). The range of pressure was 15.68 to 27.85 kPa and the average pressure applied was 20.4 kPa. Integrity of intervention: good
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time point: first 30 seconds post‐procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: trial reported that no adverse events occurred
Notes	Key conclusions of study authors: the pain score during the heel prick was significantly lower with the prior manual pressure than with the heel prick that was performed without manual pressure Study funding source: not reported Author COI statement: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
1) Selection bias (random sequence generation)	High risk	The method described is analogous to the high‐risk sequence generation methods that Cochrane describes.
2) Selection bias (allocation concealment)	High risk	The method described is analogous to the high‐risk allocation concealment method of "Alternation or rotation".
3a) Performance bias (blinding of personnel)	High risk	All interventions and data collection were completed by the researcher who was unblinded.
3b) Detection bias (blinding of outcome assessors)	High risk	All interventions and data collection were completed by the researcher who was unblinded.
4) Attrition bias (incomplete outcome data)	Low risk	No missing outcome data.
5) Selective reporting (reporting bias)	Unclear risk	No clinical trial number was noted so it was not possible to assess whether planned outcomes aligned with reported outcomes.
6) Other bias	Low risk	No evidence of other potential sources of bias.
Total	High risk	As per Sterne 2019, if one or more domain falls under high risk then the overall rating is high risk.

ATP: adenonsine triphosphate; CHEOPS: Children's Hospital of Eastern Ontario Pain Scale; COI: conflicts of interest; CONSORT: Consolidated Standards of Reporting Trials; DAN: Douleur Aiguë Nouveau‐né; DPT: diphtheria, pertussis and tetanus; EBI: environmental and behavioural interventions; EDIN: Échelle Douleur Inconfort Nouveau‐Né (neonatal pain and discomfort scale); EMLA: Eutectic Mixture of Local Anaesthetics; ER: emergency room; ES: endotracheal suctioning; FLACC: Face Legs Arms Cry Consolability Scale; FRSQ: Fonds de Recherche en Santé du Québec; FT: facilitated tucking; GA: gestational age; IM: intramuscular; IQR: interquartile range; KMC: kangaroo mother care; MAISD: Measure of Adult and Infant Soothing and Distress; MBPS: Modified Behavioral Pain Scale; N/A: not available; NESAP: noninvasive electrical stimulation of acupuncture points; NFCS: Neonatal Facial Coding System; NICU: neonatal intensive care unit; NIDCAP: Newborn Individualized Developmental Care and Assessment Program; NIH PCRC: National Institute of Health Pediatric Clinical Research Center; NIH: National Institute of Health; NINR: National Institute of Nursing Research; NIPS: Neonatal Infant Pain Scale; NNS: non‐nutritive sucking; N‐PASS: Neonatal Pain, Agitation, and Sedation Scale; OGT: oral‐gastric tube; OS: oral sucrose; PCV: pneumococcal conjugate vaccine; PHRC: French Government Funding Program for Clinical Research; PI: principal investigator; PIPP: Premature Infant Pain Profile; PMA: postmenstrual age; RA: research assistant; RCT: randomised controlled trial; SD: standard deviation; SES: socioeconomic status; SESEP: Société d'Études et de Soins pour les Enfants Paralysés et Malformés; TcB: transcutaneous bilirubin; VAS: visual analogue scale

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Abbasoglu 2015a	No control group/active control
Abdallah 2013	Not an RCT
Abedin 2008	Wrong age or age could not be separated
Aguirre 2008	No control group/active control
Akcan 2009	Wrong intervention
Al‐Bekaa 2003	Wrong age or age could not be separated
Alidadian 2021	Wrong intervention
Alix‐Seguin 2017	Wrong study type
Altuntas 2022	Wrong intervention
Angeles 2015	No acute painful procedure
Ashok 2017	Wrong intervention
Aslan 2022	Wrong age or age could not be separated
Aslanabadi 2008	Wrong age or age could not be separated
Atchariya 2017	Wrong age or age could not be separated
Axelin 2010	Wrong intervention
Azarmnejad 2018	No behavioural/multidimensional pain measure
Badiee 2013	No control group/active control
Bagherian 2016	Wrong age or age could not be separated
Bailey 2015	Wrong age or age could not be separated
Banga 2016	Wrong intervention
Barandouzi 2020	Wrong intervention
Bay 2020	Wrong study type
Beheshtipoor 2017	Not an RCT
Bekar 2022	Wrong intervention
Bellieni 2003	Wrong intervention
Bellieni 2007	Wrong intervention
Ben‐Pazi 2017	Wrong age or age could not be separated
Benoit 2021	Wrong intervention
Benzer 2017	Wrong intervention
Bergomi 2014	Wrong intervention
Boots 2010	Wrong intervention
Bos‐Veneman 2018	Wrong intervention
Bourdier 2019	No control group/active control
Britto 2017	Wrong intervention
Bueno 2010	No control group/active control
Byun 2018	Wrong age or age could not be separated
Calcaterra 2014	Wrong intervention
Campbell‐Yeo 2009	Wrong study type
Campbell‐Yeo 2014	No behavioural/multidimensional pain measure
Campbell‐Yeo 2019	Wrong intervention
Campos 1989	No control group/active control
Canbulat 2015	Wrong age or age could not be separated
Cardoso 2014	Wrong intervention
Carlson 2000	Wrong age or age could not be separated
Carneiro 2018	No acute painful procedure
Castral 2008	Wrong intervention
Ceylan 2018	No control group/active control
Ceylan 2021	No acute painful procedure
Chermont 2009	Wrong intervention
Choi 2016	Wrong intervention
Choudhary 2016	Wrong intervention
Cignacco 2008	No control group/active control
Collados‐Gomez 2017	Wrong intervention
Cologna 1999	No pain/pain not measured during procedure
Cong 2006	Wrong intervention
Cong 2009	No behavioural/multidimensional pain measure
Cuhaci 2020	Wrong intervention
Curry 2012	Wrong intervention
D'Agostino 2008	Wrong age or age could not be separated
Dahlquist 2002	Wrong age or age could not be separated
Dahlquist 2005	Wrong age or age could not be separated
Dar 2019	Wrong intervention
De Bernardo 2019	Wrong intervention
De Jong 2012	No acute painful procedure
de Sousa 2008	Wrong intervention
Desjardins 2016	Wrong intervention
Diego 2009	No control group/active control
Dilen 2010	Wrong intervention
Dolgun 2017	No control group/active control
Dos Santos 2015	Not an RCT
Downey 2012	Wrong age or age could not be separated
Drago 2009	Wrong age or age could not be separated
Duncan 2004	No pain/pain not measured during procedure
Dur 2018	Wrong intervention
El‐Naggar 2010	Wrong intervention
Elmoneim 2020	No pain/pain not measured during procedure
Erkul 2017	Wrong intervention
Erkut 2021	No control group/active control
Evelien 2020	Wrong study type
Fallah 2017	No control group/active control
Favara‐Scacco 2001	Wrong age or age could not be separated
Felt 2000	Not an RCT
Ferber 2008	Wrong intervention
Ferdosian 2020	Wrong intervention
Franck 2011	No behavioural/multidimensional pain measure
Gad 2019	Wrong intervention
Gad 2021	Wrong article type
Gan 2020	No control group/active control
Gao 2020	Wrong intervention
Gautheyrou 2018	No acute painful procedure
Gedaly‐Duff 1992	Wrong age or age could not be separated
Gezginci 2021	Wrong age or age could not be separated
Gold 2006	Wrong age or age could not be separated
Gonzalez 1989	Wrong age or age could not be separated
Gonzalez 1993	Wrong age or age could not be separated
Goubet 2003	No control group/active control
Goubet 2007	No control group/active control
Gouin 2018	Wrong intervention
Gray 2000	Wrong intervention
Gray 2012	No control group/active control
Greenberg 1997	Wrong study type
Grunau 2004	No control group/active control
Ha 2013	No acute painful procedure
Haixia 2015	Wrong intervention
Haixia 2021	No control group/active control
Hanson 2010	Wrong intervention
Harrison 2000	No acute painful procedure
Harrison 2015	Wrong intervention
Harrison 2020	No acute painful procedure
Hatem 2006	Wrong age or age could not be separated
Hawkins 1998	Wrong age or age could not be separated
He 2010	Wrong intervention
Heden 2009	Wrong age or age could not be separated
Hoffman 2006	Wrong age or age could not be separated
Holsti 2005	Wrong intervention
Howard 2019	Wrong intervention
Hsieh 2017	Wrong intervention
Hsu 1995	No pain/pain not measured during procedure
Huang 1999	No pain/pain not measured during procedure
Huang 2004	No control group/active control
Hurley 2020	Wrong intervention
Ida 2008	Wrong age or age could not be separated
Ipp 2007	Wrong intervention
Ipp 2009	No control group/active control
Jackson 2008	Wrong age or age could not be separated
Jie 2017	Wrong intervention
Jiemin 2015	Wrong intervention
Jo 2007	Wrong age or age could not be separated
Johnston 2003	Wrong intervention
Johnston 2007b	No pain/pain not measured during procedure
Johnston 2008a	No control group/active control
Johnston 2009	No control group/active control
Johnston 2012	Wrong intervention
Karimi 2022	Wrong intervention
Kashaninia 2008	Wrong intervention
Kassab 2020	Wrong intervention
Kaur 2016	Wrong age or age could not be separated
Kazmi 2020	Wrong intervention
Khajeh 2019	No acute painful procedure
Khalid 2019	Wrong intervention
Khanjari 2021	Wrong age or age could not be separated
Kivijärvi 2008	Wrong age or age could not be separated
Koivusalo 2009	Wrong age or age could not be separated
Konar 2021	Wrong intervention
Kostandy 2008	Wrong intervention
Kucukoglu 2015a	Wrong intervention
Kumari 2021	Wrong age or age could not be separated
Kurdahi 2017	Wrong intervention
Lago 2019	Wrong intervention
Lamy 2019	Wrong intervention
Lan 2018	No pain/pain not measured during procedure
Lavin 2019	No pain/pain not measured during procedure
Lavoie 2015	No acute painful procedure
Leclair 2007	No pain/pain not measured during procedure
Lee 2021	Wrong age or age could not be separated
Leite 2009	Wrong intervention
Li 2007	Wrong age or age could not be separated
Lima 2013	Not an RCT
Lima 2017	No control group/active control
Liu 2015	Wrong intervention
Lucas 2019	Wrong age or age could not be separated
Ludington‐Hoe 2005	No control group/active control
Luthy 2013	No behavioural/multidimensional pain measure
MacLaren 2005	Wrong age or age could not be separated
Maghsoudi 2015	Wrong age or age could not be separated
Mandee 2021	Wrong intervention
Manne 1990	Wrong age or age could not be separated
Marchisotti 2007	Wrong age or age could not be separated
Marec‐Berard 2009	Wrong age or age could not be separated
Marin Gabriel 2010	No pain/pain not measured during procedure
Marofi 2015	No behavioural/multidimensional pain measure
Marofi 2018	Wrong intervention
Maya‐Enero 2022	Wrong intervention
McCarthy 2010	Wrong age or age could not be separated
McNair 2017	No pain/pain not measured during procedure
Melo 2021	Wrong intervention
Mercer 2013	Wrong age or age could not be separated
Metreş 2019	Wrong intervention
Michel 2008	Wrong age or age could not be separated
Miguez‐Navarro 2016	Wrong age or age could not be separated
Miles 2006	Wrong intervention
Mirlashari 2021	Wrong intervention
Modanloo 2021	No pain/pain not measured during procedure
Moline 2020	Wrong age or age could not be separated
Montero‐Ruiz 2020	Wrong intervention
Morelius 2009	Not an RCT
Mosayebi 2014	Wrong intervention
Mucignat 2004	Wrong intervention
Naik 2021	Wrong intervention
Nandini 2021	Not an RCT
Nang 2021	Wrong study type
Napiórkowska‐Orkisz 2022	No control group/active control
Nayak 2020	Wrong intervention
Neshat 2016	No behavioural/multidimensional pain measure
Newell 2018	Wrong age or age could not be separated
Nieto 2018	Not an RCT
Nimbalkar 2020	Wrong intervention
Obeidat 2015	Wrong intervention
Okan 2010	Wrong intervention
Olsson 2019	Wrong intervention
Ozdogan 2010	Wrong intervention
Pandita 2018	No control group/active control
Pasha 2017	No control group/active control
Passariello 2020	Wrong intervention
Patel 2022	No control group/active control
Phipps 2005	Wrong age or age could not be separated
Potts 2017	Wrong age or age could not be separated
Pouraboli 2019	Wrong intervention
Rackini 2020	Wrong age or age could not be separated
Ramar 2019	Wrong intervention
Rattaz 2005	No control group/active control
Rawal 2018	Wrong intervention
Reichel 2007	Wrong age or age could not be separated
Ricero‐Luistro 2020	Wrong study type
Richey 2022	Wrong age or age could not be separated
Rioualen 2018	Wrong intervention
Rodrigues 2017	Wrong intervention
Rosali 2015	Wrong intervention
Rossi 2017	Wrong intervention
Saeidi 2011	Wrong intervention
SalawatiGhasemi 2021	Wrong age or age could not be separated
Salmani 2017	No behavioural/multidimensional pain measure
Salmon 2006	Wrong age or age could not be separated
Sağlık 2019	Wrong age or age could not be separated
Sen 2020	Wrong intervention
Shabani 2016	Wrong intervention
Shah 2006	No control group/active control
Shah 2017	Wrong intervention
Shapiro 2007	Wrong age or age could not be separated
Sharara‐Chami 2022	Wrong intervention
Shen 2015	Wrong intervention
Shoghi 2021	Wrong age or age could not be separated
Shukla 2018	Wrong intervention
Shukla 2018a	Wrong intervention
Shukla 2021	Wrong intervention
Singh 2012	Not an RCT
Singh 2016	No control group/active control
Sinkey 2015	Wrong intervention
Slater 2010	Wrong intervention
Slifer 2009	Wrong age or age could not be separated
Sola 2017	No acute painful procedure
Sorrentino 2017	Wrong intervention
Sparks 1998	Wrong age or age could not be separated
Sparks 2007	Wrong age or age could not be separated
Sri 2017	Wrong age or age could not be separated
Stadler 2021	No control group/active control
Stevens 2005	Wrong intervention
Stevens 2014	Not an RCT
Stevens 2016	Wrong intervention
Stevens 2018	Wrong intervention
Sujatha 2018	Wrong intervention
Suksumek 2017	Wrong intervention
Sundar 2016	Wrong intervention
Sundararajan 2007	Wrong age or age could not be separated
Sunitha 2015	No acute painful procedure
Sureka 2016	No acute painful procedure
Susam 2018	Wrong age or age could not be separated
Taddio 2014a	No acute painful procedure
Taddio 2015a	No pain/pain not measured during procedure
Taddio 2016	No acute painful procedure
Taddio 2018	No pain/pain not measured during procedure
Tanabe 2002	Wrong age or age could not be separated
Tang 2018	Wrong intervention
Taplak 2017	Wrong intervention
Tasci 2020	No control group/active control
Tavlar 2022	No control group/active control
Tekgündüz 2019	Wrong intervention
Teunkens 2018	Wrong intervention
Toshiyuki 2016	No acute painful procedure
Tudor‐Drobjewski 2018	No acute painful procedure
Ucar 2014	No acute painful procedure
Ueki 2021	Wrong age or age could not be separated
Uematsu 2019	Wrong intervention
Ullan 2018	No acute painful procedure
Ullsten 2017	Wrong intervention
Usclade 2022	No control group/active control
Usichenko 2016	Wrong age or age could not be separated
van der Heijden 2018	Wrong intervention
Varghese 2018	Not an RCT
Varghese 2020	Wrong age or age could not be separated
Veronesi 2016	No acute painful procedure
Viana 2020	Wrong intervention
Viggiano 2021	Wrong intervention
Vignochi 2010	No pain/pain not measured during procedure
Vitale 2017	Wrong intervention
Vivancos 2010	Not an RCT
Wang 2021	Wrong intervention
Weissman 2009	Not an RCT
Wen 2018	Wrong intervention
Williams 2019	No behavioural/multidimensional pain measure
Wirth 2014	No acute painful procedure
Wisdorf‐Houtkooper 1997	Wrong intervention
Wu 2020	Wrong intervention
Wu 2021	Wrong intervention
Xin 2019	No acute painful procedure
Yan 2015	No acute painful procedure
Yildirim 2018	Wrong age or age could not be separated
Yin 2017	No control group/active control
Yin 2022	No behavioural/multidimensional pain measure
Zeltzer 1991	Wrong age or age could not be separated
Zeraati 2015	No behavioural/multidimensional pain measure
Zhang 2018	Wrong age or age could not be separated
Zun 2012	Wrong study type
Zurita‐Cruz 2017	Wrong intervention
Çakıcı 2020	Wrong intervention

RCT: randomised controlled trial

Characteristics of studies awaiting classification [ordered by study ID]

Arditi 2006.

Methods	Study design: between‐group Total study duration: not clear
Participants	Total number: 62 in study 1 and a new group of 62 in study 2 Setting: nursery Diagnostic criteria: healthy newborns with singleton vaginal birth Gestational age: Study 1: Mean age was 51.8 hours (SD 7.3, range  41 to 67) Mean gestational age at birth was 39.8 weeks (SD  1.1, range  37 to 43) Sex: Males: 29 Females: 33   Study 2: Mean gestational age at birth was 39.7 weeks (SD  1.11, range 37 to 42) and mean age at testing was 53.5 hours (SD  8.29, range 40 to 76) Sex: Males: 33 Females: 29 Country: Israel
Interventions	Total number of intervention groups: 1 Control group description: infants in the control group were in the bassinet Total number of participants randomised to each of the groups: 62  Specific intervention: infants were held by a female research assistant with the infant’s head in the crook of the left arm so that the left leg was available for the heel‐prick sampling
Outcomes	NFCS High score = more pain Continuous Adverse events: no
Notes	Key conclusions of study authors: human contact had no effect on the newborns’ pain response Study funding source: Israel Science Foundation (grant # 01/941) and the March of Dimes Foundation (#12‐ FY04‐50) Author COI statement: no

AvanAntepli 2021.

Methods	Study design: between‐group Total study duration: 3 months
Participants	Total number: 56 Setting: primary care clinic Diagnostic criteria: 0‐ to 28‐day‐old full‐term newborns Gestational age: gestational ages of 38 weeks and older Experimental group: mean ± SD 38.7 ± 0.6 Control: 38.8 ± 0.8   Sex: Males: 28 Females: 28 Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: no interventions were provided to the newborns in the control group Total number of participants randomised to each of the groups: 28 Specific intervention: a vibrating device (Gray mini vibrator, 8.5 cm, 92 Hz to 100 Hz) was applied below the knee of the extremity from which the blood was to be collected. This was done approximately 30 seconds before the commencement of each heel lance procedure. The vibration was continued throughout the heel lance procedure and was stopped immediately after the procedure was over.
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: pain scores provided by the nurses and specialists before the heel lance procedure 15 to 20 seconds and 5 minutes after the procedure Upper limit: 7 Lower limit: 0 High score = more pain Continuous   Adverse events: trial did not report
Notes	Key conclusions of study authors: the pain scores were significantly lower at 15 to 20 seconds and 5 minutes after the heel lance procedures in the group that received vibrations than in the control group Study funding source: Gaziantep University; it is a master’s thesis Author COI statement: no conflicts of interest

Bagheri 2020.

Methods	Study design: between‐group Total study duration: N/A
Participants	Total number: 64 Setting: NICU ward of a medical centre Diagnostic criteria: preterm infants who needed insertion of umbilical vein catheter Gestational age: 30th to 37th week of gestational age Experimental: 32  Control: 32  Males: not reported  Females: not reported Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: all of the treatment steps were identical for the infants in the control group except who were not given massage Total number of participants randomised to each of the groups: 32 Specific intervention: in the experimental group, the infants were given 15‐minute massage (Field technique) consisting of 3 x 5‐minute phases
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: first 30 seconds post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors: pain scores were significantly lower in the experimental group compared to the control group. Massage therapy can be effective in management and pain relief.  Study funding source: Mashhad University of Medical Sciences Author COI statement: no conflict of interest

Bertrand 2019.

Methods	Could not retrieve data as full‐text could not be accessed.
Participants	Could not retrieve data as full‐text could not be accessed.
Interventions	Could not retrieve data as full‐text could not be accessed.
Outcomes	Could not retrieve data as full‐text could not be accessed.
Notes	Could not retrieve data as full‐text could not be accessed.

Cakirli 2021.

Methods	Study design: between‐group Total study duration: 4 months
Participants	Total number: 90 Setting: hospital Diagnostic criteria: being born at term Gestational age: 38 to 40 weeks Mean: 39 SD: N/A Sex: Males: 13 Females: 17 Country: Turkey
Interventions	Total number of intervention groups: 2 Control group description: a loose swaddling was applied to the newborns in all groups as a pain control intervention Total number of participants randomised to each of the groups: 30 Specific intervention:  Smell of own mother’s milk: gauze placed inside container of milk and placed 15 cm from the baby's nose starting at 2 minutes before procedure and continuing post‐procedure until infants returned to baseline Smell of another mother’s milk: gauze placed inside container of milk and placed 15 cm from the baby's nose starting at 2 minutes before procedure and continuing post‐procedure until infants returned to baseline
Outcomes	Pain outcomes, scale limits, and time points when measured N‐PASS Time points: not found (possibly 2 minutes after needle) Upper limit: 11 Lower limit: 0 High score = more pain Continuous Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors: the group exposed to their own mother’s milk had a significantly lower N‐PASS mean score compared to other groups and the group which was exposed to another mother’s milk was observed to have a significantly lower N‐PASS mean score compared to the control group. The most effective method for reducing pain in newborns was the smell of their own mother’s breast milk. Study funding source: did not receive any specific grant from funding Author COI statement: no potential conflicts of interest

Chen 2021.

Methods	Study design: between‐group Total study duration: 2 to 3 months
Participants	Total number: 116 Setting: hospital Diagnostic criteria: full‐term at birth Gestational age: gestational age of at least 32 weeks Control: 38.5 ± 1.7 Experimental: 38.5 ± 1.6 Sex: Males: 26 Females: 32 Country: China
Interventions	Total number of intervention groups: 1 Control group description: control group received routine care which consisted of swaddling using a blanket Total number of participants randomised to each of the groups: 58 Specific intervention: on top of routine care, maternal voice was played to the neonate by a voice recorder on one side of the head at a distance of 20 cm from the neonate's ear. The sound was set between 50 and 60 decibels by using a decibel meter. Maternal voice was played for 2 minutes before the procedure and until the neonate's heart rate and oxygen saturation returned to baseline.
Outcomes	NIAPAS A 2, 3, or 4‐point scale (0, 1; or 0, 1, 2; or 0, 1, 2, 3) for a possible total score of 18 High score = more pain Continuous Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors:  Behavioural and physiological indicator scores in the maternal voice group were significantly lower compared with those in the routine care group Study funding source: funds from School of Nursing, Peking Union Medical College Author COI statement: none

Cignacco 2012.

Methods	Study design: between‐group Total study duration: 12 months
Participants	Total number: 71 Setting: NICUs of 3 university hospitals Diagnostic criteria: first 14 days of their NICU; born between 24 and 32 weeks of gestation and anticipated clinical need for at least 5 routine capillary blood samples within 2 weeks after birth Gestational age: preterm infants between 24 and 32 weeks of gestation Mean: 29.2  SD: (1.8) Sex: Males: 39 Females: 32 Country: Switzerland
Interventions	Total number of intervention groups: 1 Control group description: sucrose was administrated orally ∼2 minutes before the heel stick Total number of participants randomised to each of the groups: 24 Specific intervention: facilitated tucking (FT) was started at the beginning of the baseline phase, and the infant was “tucked” through all 3 phases. The FT was started at the beginning of the baseline phase and sucrose was given 2 minutes before the heel stick.
Outcomes	Pain outcomes, scale limits, and time points when measured BPSN Time points: 3 phases (baseline, heel stick, 3 minutes after the heel stick) 9 items: 3 physiologic (heart rate, respiratory rate, and oxygen saturation) and 6 behavioural (grimacing, body movements, crying, skin colour, sleeping patterns, consolation) items Each item was scored on a 3‐point scale (0 to 3 points) High score = more pain Continuous Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors: FT alone was significantly less effective in relieving repeated procedural pain than sucrose (0.2 mL/kg). FT in combination with sucrose seemed to have added value in the recovery phase with lower pain scores compared with both the single‐treatment groups. Study funding source: Swiss National Science Foundation and  private foundations Author COI statement: not mentioned

Deniz 2021.

Methods	Study design: between‐group  Total study duration: ~5 months
Participants	Total number: 105 Setting: hospital Diagnostic criteria: delivered by caesarean section Gestational age: 37 to 41 weeks Mean: n/a SD: n/a Sex: Males: 16, 18, 18 (control) Females: 19, 17, 17 (control) Country: Turkey
Interventions	Total number of intervention groups: 2 Control group description: no pre‐application was given to the babies in the control group as a routine procedure Total number of participants randomised to each of the groups: 35 Specific intervention:  Acupressure: the acupressure session involved relaxation for the first 1 minute; 90 seconds of acupressure was applied to each point (Ki3 and St36 points). Acupressure was performed for each newborn for only 1 session that lasted 7 minutes. Foot reflexology: the foot reflexology session involved relaxation for the first 1 minute; the application was performed in the remaining 6 minutes as advised in the literature. Foot reflexology was performed for each newborn for only 1 session that lasted 7 minutes.
Outcomes	Pain outcomes, scale limits, and time points when measured N‐PASS Time points: 3, pre‐, during and post‐intervention of heel prick Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors: acupressure and foot reflexology administered during heel lancing in newborns are effective methods for reducing pain Study funding source: no financial support Author COI statement: no potential conflicts of interest

Dorneles da Silveira 2021.

Methods	Study design: cross‐over Total study duration: March 2014 to May 2015
Participants	Total number: 34 Setting: hospital Diagnostic criteria: premature infants submitted to heel puncture for blood glucose monitoring Gestational age: 34.4 ± 11.3 Sex Males: 50% Females: 50% Country: Brazil
Interventions	Total number of intervention groups: 2 Control group description: 25% glucose was offered in a 1 mL volume directly on the anterior part of the tongue with a 1 mL syringe without a needle, 2 minutes before puncture Total number of participants randomised to each of the groups: 34 Specific intervention: 25% glucose + NNS was offered 2 minutes before puncture, with glucose being offered first and then NNS. NNS was offered with a gloved little finger 2 minutes before and during puncture.
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: 30 seconds before the intervention, called the baseline period and for 5 minutes after puncture Upper limit: 21 Lower limit: 0 High score = more pain Continuous   Adverse event episode (%)  Choking 11.8 Bradycardia 2.9 Oxygen desaturation 8.8 Tachycardia 2.9 Nausea 2.9
Notes	Key conclusions of study authors: effect of isolated and combined interventions showed that, when offered in combination, preterm infants were able to return to baseline parameters more quickly Study funding source: not found Author COI statement: not found

Erdogan 2022.

Methods	Study design: between‐group Total study duration: not found
Participants	Total number: 120 Setting: NICU hospital Diagnostic criteria: infants ‡ 37 weeks old Gestational age: ~38 Mean: ~38 SD: +/– 1.0, 1.3, 1.3, 1.1 for each group Sex Female: taste + smell (15), taste (15), smell (16), control group (10) Total n = 30 in each group   Country: Turkey
Interventions	Total number of intervention groups: 3 Control group description: no intervention given  Total number of participants randomised to each of the groups: 30 Specific intervention:  Group 1: 5 mL of milk was dripped into the infant’s mouth to give the infant a taste of the breast milk Group 2: 5 mL of milk‐soaked cotton was placed at a 5 cm to 10 cm distance to the infant’s nose for the infant to smell it Group 3: the procedures in both groups 1 and 3 were performed
Outcomes	Pain outcomes, scale limits, and time points when measured N‐PASS Time points: 3: before blood draw, during, after Upper limit: 11 Lower limit: 0 High score = more pain Continuous Adverse events: trial did not report on adverse events
Notes	Key conclusions of study authors: breast milk smell and taste are recommended for use as a nursing intervention to reduce pain, decrease the HR, and increase SpO2 in the NICU during painful procedures Study funding source: no funding Author COI statement: no competing financial interests exist

Güngör 2021.

Methods	Study design: between‐group Total study duration: 5 months
Participants	Total number: 96 Setting: hospital, family health centres Diagnostic criteria: infants 2 to 6 months of age Gestational age: 38.5 ± 1.0 (control), 38.6 ± 0.9 (heat), 38.1 ± 0.9 (cold) Sex Males: 17 (control), 16 (heat), 16 (cold) Females: 16 (control), 15 (heat), 17 (cold) Country: Turkey
Interventions	Total number of intervention groups: 2 Control group description: control group did not undergo any intervention before the vaccination Total number of participants randomised to each of the groups: 32 Specific intervention:  Heat application: heat thermogel compresses were held in boiled hot water for 15 minutes before the procedure. Heat compresses were covered with a sheath before the procedure in order to prevent direct contact with the skin. Heat was applied locally to the area to be vaccinated for 2 minutes. Cold application: cold thermogel compresses used in the study were kept in the freezer of the refrigerators found in the family health centres. Cold compresses were covered with a sheath before the procedure in order to prevent direct contact with the skin. Cold was applied locally to the area to be vaccinated for 2 minutes.
Outcomes	Pain outcomes, scale limits, and time points when measured FLACC Time points: pre, during, post intervention of vaccine Upper limit: 10 Lower limit: 0 High score = more pain Continuous Adverse events: non reported
Notes	Key conclusions of study authors: local cold and heat application methods applied to the vaccination area before a pneumococcal vaccine reduced vaccine‐associated pain in the infants, and the application of cold was more effective than heat Study funding source: not found Author COI statement: none declared

Hoseini 2020.

Methods	Study design: between‐group Total study duration: not found
Participants	Total number: 76 Setting: teaching hospitals Diagnostic criteria: full‐term Gestational age: Mean: 31.9 ± 2.2 (intervention), 31.7 ± 2.6 control  Sex Males: 17 (intervention), 20 control  Females: 21 (intervention), 18 control  Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: not given intervention Total number of participants randomised to each of the groups: 38 Specific intervention: earmuffs were used 4 hours before orogastric tube insertion and eye shields were used 2 hours before the procedure
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: 2 to 3 minutes pre‐intervention, during intervention, 30 seconds post‐procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous   Adverse events: not reported
Notes	Key conclusions of study authors: eye shields and earmuffs could be used before orogastric tube insertion in order to relieve pain Study funding source: not found Author COI statement: none reported

Inal 2021.

Methods	Study design: between‐group Total study duration: not found
Participants	Total number: 105 Setting: NICU hospital Diagnostic criteria: full‐term neonates (38 to 42 weeks of gestation) Gestational age: Mean ± SD: 39.0 ± 0.98 Sex Males: 52 Females: 53 Country: Turkey
Interventions	Total number of intervention groups: 2 Control group description: the heel stick procedures were conducted using the standard method and the neonates received no interventions during the procedures Total number of participants randomised to each of the groups: 35 Specific intervention:  Swaddling: neonates were placed in the supine position on a blanket. The legs were wrapped in the flexion and abduction position. The arms of the neonates were placed close to their torso with both hands, without restraining limb movements. Swaddling was performed 1 minute before the heel stick procedure and continued 3 minutes after the procedure. Maternal holding: neonates were held in their mothers’ lap while their mothers were seated reclining on a comfortable chair. Neonates remained clothed in their mothers’ lap during the heel stick procedure. Holding continued for a minimum of 3 minutes during and after the procedure.
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: from the beginning of the heel stick until the end of the recovery period; 3 minutes total Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse events: not reported
Notes	Key conclusions of study authors: both swaddling and maternal holding are effective at reducing pain felt during heel stick procedures, but mother holding may be preferred as a priority rather than swaddling Study funding source: no specific funding Author COI statement: no conflicts of interest

Jin 2018.

Methods	Study design: between‐group Total study duration: 8 months
Participants	Total number: 200 Setting: hospital Diagnostic criteria: newborns with a gestational age of < 35 weeks Gestational age: Gestational age (weeks) 38 ± 1 Sex Male: 17 (control), 20 (sterile), 16 (dex), 23 (pacifier) Country: Korea
Interventions	Total number of intervention groups: 3 Control group description: nothing given Total number of participants randomised to each of the groups: 33 to 39 (50 initially) Specific intervention: sterile water (n = 35), 25% dextrose (n = 35), and pacifier group (n = 39)
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS and NFCS Time points: before, during, and 2 minutes after Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse events: not found
Notes	Key conclusions of study authors: oral administration of 25% dextrose solution for pain control during the newborn screening test led to a significantly lower salivary cortisol level than the use of sterile water or a pacifier. However, no difference in pain scores was found among groups. Study funding source: Gangneung Asan Hospital Biomedical Research Center Promotion fund Author COI statement: no conflict

Karimi 2019.

Methods	Study design: between‐group Total study duration: not found
Participants	Total number: 34 Setting: hospital Diagnostic criteria: birth age of less than 37 weeks Gestational age: Control 35.3 ± 3.3 Swaddling 34.0 ± 3.0 Sex Males: 39 Females: 29 Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: no intervention given  Total number of participants randomised to each of the groups: 17 Specific intervention: swaddling group were swaddled by bed sheet 10 minutes before blood sampling until 2 minutes after it
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: intervals of 30 seconds before, during, and 30, 60, 90, and 120 seconds after blood sampling Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse events: not reported
Notes	Key conclusions of study authors: swaddling can reduce the risk of blood sampling in preterm infants and keep their vital signs in a better status Study funding source: not reported  Author COI statement: declared none

Kassab 2020b.

Methods	Study design: between‐group
Participants	Total number: 135 Diagnostic criteria: neonates who are 38 to 40 gestational weeks
Interventions	Total number of intervention groups: 2 Control group description: in the control group, no intervention was received by the neonates, and routinely applied procedures of glucose heel check were performed Total number of participants randomised to each of the groups: 45 participants randomised to each of the groups (45 in the oral dextrose water D10%W group, 45 in the FTP group, and 45 in the control group) Specific intervention:  Oral dextrose water D10%W: 2 mL of dextrose DW 10% was given to each neonate orally 15 seconds before and during heel lance procedure by a nurse Facilitated tucking by parents (FTP): parents performed facilitated tucking for at least 2 minutes before and during heel lance procedure and continued for another 2 minutes after procedure
Outcomes	Pain outcomes, scale limits, and time points when measured MBPS Time points: 3: before, during, and after the injection Upper limit: 10 Lower limit: 0 High score = more pain Continuous
Notes	Key conclusions of study authors: both dextrose water and facilitated tucking were found to be effective in reducing behavioural scores compared to control group

Kim 2022.

Methods	Study design: between‐group Total study duration: 7 months
Participants	Total number: 48 Setting: hospital Diagnostic criteria: gestational age, within 28 to 37 weeks premature infants Gestational age: Mean: 33.3 SD: ± 2.3 Sex Males: 19 Females: 25 Country: Korea
Interventions	Total number of intervention groups: 1 Control group description: routine care Total number of participants randomised to each of the groups: 24 (22 analysed) Specific intervention: cotton swab soaked in breast milk after which the milk was gently applied to the infants’ nose by moving the cotton swab slowly from one nostril to the other in 1 second. This was repeated 3 times. The exposure time was 30 seconds.
Outcomes	Pain outcomes, scale limits, and time points when measured PIPS Time points: baseline and 30 seconds after pain Upper limit: 12 Lower limit: 0 High score = more pain Continuous Adverse events: not reported
Notes	Key conclusions of study authors: breast milk odour may reduce pain in premature infants during heel stick tests Study funding source: no competing financial interests exist Author COI statement: this research was funded by Seoul Nurses Association Hanmaeum Scholarship in 2018

Kucukoglu 2016.

Methods	Study design: between‐group Total study duration: 3 months
Participants	Total number: 75 Setting: university hospital Diagnostic criteria: premature infants Gestational age: Gestational age at birth, weeks 31.8 ± 3.3 (study group), 31.3 ± 2.5 (control) Sex Males: 33 Females: 42 Country: Turkey
Interventions	Total number of intervention groups: 78 Control group description: routine clinical practice, without any intervention Total number of participants randomised to each of the groups: 35 in the study group and 40 in the control group Specific intervention: white noise using an MP3 was played for 1 minute before the hepatitis B vaccination was performed, and continued until 1 minute after the vaccination
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: 1 minute before vaccination was performed, and until 1 minute after the vaccination Upper limit: 21 Lower limit: 0 High score = more pain Continuous
Notes	Key conclusions of study authors: white noise was found to be effective for this sample; however, there is a dire need for extensive research on white noise and its use with this vulnerable population Study funding source: not reported Author COI statement: not reported

Lan 2021.

Methods	Study design: between‐group Total study duration: not found
Participants	Total number: 120 Setting: hospital Diagnostic criteria: premature infants Gestational age: GA of 39.1 ± 1.0 weeks Sex Females and males was similar (50.8% and 49.2%, respectively) Country: Taiwan
Interventions	Total number of intervention groups: 3 Control group description: a female nurse gently touched the head (GT) and spoke softly to comfort the newborn (VC) during and after the heel stick Total number of participants randomised to each of the groups: 40 Specific intervention:  Condition 2 (breast milk odour): 3 minutes prior to the heel stick to the 5th minute of recovery, a cotton ball with the breast milk was put near the newborn’s nostrils for breast milk odour. GT and verbal comfort was also offered.  Condition 3 (breast milk odour + taste): on top of routine care (GT and verbal comfort), and exposure to breast milk odour, 3 mL of mother’s breast milk were fed slowly through syringe dripping to the newborn’s mouth 2 minutes before and during the heel stick
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: 5 minutes before heel stick without stimuli (baseline); during heel stick; recovery (3 total) Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse events: not reported
Notes	Key conclusions of study authors: compared with the routine care, provision of the odour and taste of breast milk reduce NIPS scores during heel sticks and during recovery Study funding source: a research grant from the Ministry of Science and Technology Author COI statement: the authors declare no conflict of interest

Lin 2022.

Methods	Study design: between‐group Total study duration: 12 months (September 2017 to September 2018)
Participants	Total number: 114 Setting: hospital Diagnostic criteria: full‐term Gestational age: Mean GA of 39.1 ± 1.0 weeks Sex Males: 57 Females: 57 Country: Taiwan
Interventions	Total number of intervention groups: 3 Control group description: the neonates in the control group received only GT and comforting voice from the intervener, with one hand of the intervener placed on the head of the neonate during and after heel pricks Total number of participants randomised to each of the groups: 38 Specific intervention:  Breast milk odour (BMO): neonates received the control condition in addition to BMO, provided by placing a cotton ball soaked with 2.5 mL of breast milk (BM) 3 cm from the neonates nostril 2 minutes prior to the heel prick, and was left in place until the collection of the blood sample was completed. BMO + breast milk taste (BMT): neonates received the control condition, BMO as for described for Group 2, and BMT delivered by syringe feeding. The intervener slowly provided 2.5 mL of BM orally to the neonate from a syringe 1 minute before and during the heel prick procedure.
Outcomes	Pain outcomes, scale limits, and time points when measured The time required for neonates to stop crying (crying cessation) was used a behavioural measure of neonatal procedural pain from baseline to 6 minutes after the heel prick High score = more pain Continuous Adverse events: not reported
Notes	Key conclusions of study authors: neonates receiving either BMO or BMO plus BMTSF would have reductions in changes of heart rates and SpO2 and shorter time to crying cessation during and after a heel prick procedure compared with neonates receiving GT and human voice interventions Study funding source: research grant from the Ministry of Science and Technology Author COI statement: the authors declare no conflict of interest

Min‐Kyung 2011.

Methods	Study design: between‐group Total study duration: unclear due to language barriers
Participants	Total number: 132 Setting: hospital Diagnostic criteria: healthy full‐term newborns Gestational age: Mean: 39.1 (glucose), 39.0 (non‐nutritive sucking), 38.5 (glucose + NNS), 38.7 (control) SD: 1.1 (glucose), 1.2 (non‐nutritive sucking), 1.3 (glucose + NNS), 1.5 (control) Sex: Males: 66 Females: 66 Country: Korea
Interventions	Total number of intervention groups: 3 Control group description: 2 mL distilled water during intramuscular injection of hepatitis B vaccine Total number of participants randomised to each of the groups: glucose = 31; non‐nutritive sucking = 33; glucose + NNS = 29; control = 39 Specific intervention: glucose, non‐nutritive sucking, glucose + NNS Intervention details: glucose: 2 mL 25% dextrose during intramuscular injection of hepatitis B vaccine. Non‐nutritive sucking: pacifier during intramuscular injection of hepatitis B vaccine. Glucose + NNS: pacifier coating with 25% dextrose during intramuscular injection of hepatitis B vaccine Integrity of intervention: unclear due to language barriers
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: unclear due to language barriers Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: no mention of adverse events (unclear due to language barriers)
Notes	Key conclusions of study authors: oral 25% dextrose is more effective than distilled water or using pacifier with or without 25% dextrose in reducing pain during intramuscular injection of hepatitis B vaccinations Study funding source: unclear due to language barriers Author COI statement: unclear due to language barriers

Parhi 2021.

Methods	Study design: between‐group Total study duration: not found
Participants	Total number: 64 Setting: NICU hospital Diagnostic criteria: preterm Gestational age: Maximum 14 (51.9%) in control group were 28 to 32 weeks of gestational age and 10 (32.3%) in experimental group were 33 to 37 weeks of gestational age of the preterm neonates Sex 15 (55.6%) male in control group and 21 (67.7%) in experimental group Country: India
Interventions	Total number of intervention groups: 1 Control group description: standard care Total number of participants randomised to each of the groups: intervention = 27, control = 31 Specific intervention: Yakson therapeutic touch and kinesthetic stimulation in addition to standard care
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP and ABSS Time points: between baselines and immediately after an invasive procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse events: not reported
Notes	Key conclusions of study authors: the intervention has significant effect on minimisation of pain during the heel prick Study funding source: received the grant from CTRI Author COI statement: "Received the grant from CTRI" as a possible COI was stated

Park 2006.

Methods	Study design: between‐participant Study duration: not available
Participants	Total number: 32 Setting: in‐hospital setting Diagnostic criteria: not available Age: not available Minimum: not available Maximum: not available Mean: not available SD: not available Sex: Males: not available Females: not available Country: Korea
Interventions	Total intervention groups: 1 Control group description: not available Total number of participants randomised to each of the groups: intervention: 16, control: 16 Specific intervention: touch/massage‐related Intervention details: not available Integrity of intervention: not available
Outcomes	Pain outcomes, scale limits, and time points when measured: NIPS Coded 5 minutes after heel stick Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse reactions: no mention of adverse events
Notes	Funding sources: not available Key conclusions of study authors: touch/massage‐related intervention reduced pain more than the control condition Study funding source: information unknown as only abstract was extracted Author COI statement: information unknown as only abstract was extracted

Pouraboli 2021.

Methods	Study design: between‐group Total study duration: not found
Participants	Total number: 60 Setting: hospital Diagnostic criteria: preterm Gestational age: aged between 27 and 37 weeks Sex Males: 29 Females: 31 Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: a single one‐time dose of morphine 0.05 mg/kg was administered intravenously. Standard care was given during the chest tube removal.  Total number of participants randomised to each of the groups: 30 Specific intervention: the nurse placed each infant in the intervention group in a facilitated tucking position as follows: holding the infant in the side‐lying, flexed, fetal‐type position by bringing the body to middle position and placing one cupped hand on the posterior side of the infant’s head. The other cupped hand is placed on the infant’s buttocks using gentle hand pressure. A single one‐time dose of morphine 0.05 mg/kg was administered intravenously.
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: NIPS scores were assigned based on observation from the videos at 5 minutes prior to chest tube removal, immediately after removal of the chest tube, and 5 minutes after chest tube removal Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse events: no adverse effects
Notes	Key conclusions of study authors: facilitated tucking combined with morphine administration can be an effective additive intervention for pain control in infants during chest tube removal Study funding source: not found Author COI statement: no conflicts of interest

Rad 2021.

Methods	Study design: between‐group Total study duration: 13 months
Participants	Total number: 90 Setting: hospital Diagnostic criteria: preterm infants, 28 to 37 weeks of gestation Gestational age: MBMO 32.9 ± 2.4  BMO 31.5 ± 2.1  Control 32.5 ± 2.4 Sex Males: 43 Females: 47 Country: Iran
Interventions	Total number of intervention groups: 2 Control group description: 2 mL of distilled water was poured on a cotton swab and placed 3 cm away from the baby’s nose. This process started 3 minutes before vaccination and continued until the vaccination was completed. Total number of participants randomised to each of the groups: 30 Specific intervention:  Maternal breast milk odour (MBMO) group: 2 mL of maternal breast milk was poured on a cotton swab and placed 3 cm away from the baby’s nose. This process started 3 minutes before vaccination and continued until the vaccination was completed. Another mother's breast milk odour (BMO) group: 2 mL of another mother's breast milk was poured on a cotton swab and placed 3 cm away from the baby’s nose. This process started 3 minutes before vaccination and continued until the vaccination was completed.
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: pre‐ post‐intervention process started 3 minutes before vaccination and continued until the vaccination was completed Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse events: not stated
Notes	Key conclusions of study authors: stimulation with MBMO is effective in reducing pain in preterm infants Study funding source: a grant from Babol University of Medical Sciences Author COI statement: the authors declare that they have no competing interests

Ren 2022.

Methods	Study design: between‐group Total study duration: 6 months
Participants	Total number: 66 Setting: hospital Diagnostic criteria: neonates Gestational age:  Gestational age was 37 to 41 weeks Mean: 39 Sex Males: 38 Females: 22 Country: China
Interventions	Total number of intervention groups: 1 Control group description: neonates in the control group were exposed to 0 dB white noise for the sham intervention Total number of participants randomised to each of the groups: (experimental group: N = 29; control group: N = 31) Specific intervention: neonates in the experimental group were exposed to 50 dB white noise from the album “Colic Baby: White Noise for Babies” using the episode “Calm Rain for Colic Baby Sleep” delivered on a continuous loop. The research assistant placed the prepared music player about 20 cm away from the head of neonates in their incubators.
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP‐R Time points: measured in 30‐second intervals during the 5 minutes after needle withdrawal, which resulted in 10 assessment time points labelled T1 to T10, respectively Upper limit: 18 Lower limit: 0 High score = more pain Continuous Adverse events: no adverse events occurred that were related to the intervention
Notes	Key conclusions of study authors: the white noise intervention did not show beneficial effects on pain‐related cortical response or pain score Study funding source: this work was supported by grants from Guangdong Nurse Association and Department of Science and Technology of Guangdong Province Author COI statement: no conflict of interest

Sapkota 2021.

Methods	Study design: between‐group Total study duration: 2 years
Participants	Total number: 92 Setting: hospital Diagnostic criteria: neonates   Age (days) mean ± SD: intervention: 1.2 ± 0.4; control: 3.5 ± 3.4 Sex Males: 48 Females: 44 Country: Nepal
Interventions	Total number of intervention groups: 1 Control group description: for the control group, no application of heat was done before heel prick procedure Total number of participants randomised to each of the groups: 46 Specific intervention: the warmer (electro thermal water bag) was applied to the heel of the baby for 3 to 5 minutes, which provided the baby’s heel with the warmth of 38°C to 40°C. After 5 minutes, the warmer was removed and the on‐duty ward sister performed heel prick procedure to collect the blood sample as required.
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: before and after heel stick, exact time not found Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse events: not found
Notes	Key conclusions of study authors: heat application prior to heel stick is effective in reducing pain in newborns Study funding source: "Nil" Author COI statement: none declared

Sarhangi 2021.

Methods	Study design: between‐group Total study duration: 6 months
Participants	Total number: 60 Setting: hospital Diagnostic criteria: full‐term Gestational age: Gestational age (week): control: 38.1 (1.0); intervention 38.0 (1.3) Sex Males: 27 Females: 33 Country: Iran
Interventions	Total number of intervention groups: 1 Control group description: neonates received routine care except for listening to the mother’s heartbeat sounds. The speaker was placed to the incubator for neonates in the control groups, but no sound was played. Total number of participants randomised to each of the groups: 30 Specific intervention: the neonates in the intervention group were lying supine during the taking of blood samples and were in a separate room far from the intensive care unit to avoid noises. The mother’s heartbeat sound was played inside the incubator using 2 small speakers placed on either side of the neonate’s head, 20 cm from the neonate’s ears with the sound threshold about 50 dB set by an audiologist. The sound was played 10 minutes before blood sampling and was continued until 10 minutes after it.
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: every 10 minutes in 3 steps, 10 minutes before the intervention, immediately after and 10 minutes after the intervention Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse events: not found
Notes	Key conclusions of study authors: mother’s heartbeat sound can be used by nurses as a non‐pharmacologic and safe intervention along with routine care in order to reduce suffering and pain in neonates undergoing invasive and painful procedures in the intensive care unit Study funding source: Research Council affiliated with Baqiyatallah University of Medical Sciences, Tehran, Iran Author COI statement: no conflict of interests

SezerEfe 2022.

Methods	Study design: between‐group Total study duration: 2 months
Participants	Total number: 50 Setting: hospital Diagnostic criteria: being gestational age 32 to 37 weeks Gestational age: Gestational age (week): GHT: 35.0 ± 1.6; control 35.3 ± 1.8 Sex Males: 25 Females: 25 Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: after the nurse took heel blood from the preterm infants in the control group, she taped the incision area and closed the incubator cover. She routinely touched the infant during this process, but this touch did not include comforting touch. There is no routine practice related to procedural pain for infants during heel lancing in the NICU. Total number of participants randomised to each of the groups: 25 Specific intervention: GHT was performed 10 minutes before the heel lancing and continued during and after the procedure; 15 minutes of GHT was applied. The researcher placed one hand on the infant’s head and the other hand on the lower abdomen covering the waist and hips of the preterm infant for 15 minutes.
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: 10 minutes before intervention, during, and 5 minutes after  Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse events: no adverse events occurred during the study period
Notes	Key conclusions of study authors: GHT may positively affect preterm infants’ pain, comfort, and physiologic parameters Study funding source: no funding Author COI statement: no conflict of interest

Sun 2020.

Methods	Study design: between‐group Total study duration: 1 year and 4 months
Participants	Total number: 82 Setting: hospital Diagnostic criteria: preterm Gestational age: Gestational age (weeks) control: 29.9 ± 0.4; intervention: 30.0 ± 0.4 Sex Males: 34 Females: 32 Country: China
Interventions	Total number of intervention groups: 1 Control group description: routine care Total number of participants randomised to each of the groups: intervention: 28; control: 38  Specific intervention: GHT was implemented from the beginning of each procedure until 10 minutes after the procedure. The anaesthesiologists placed the fingertips of the left hand above the eyebrow line with the palm touching the preterm infant’s crown. The rest of right hand and fingers rested on the infant’s upper arm, while the right thumb was on the infant’s right shoulder (midline position).
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: not found, seems like throughout the procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse events: not found
Notes	Key conclusions of study authors: gentle human touch can effectively alleviate pain during ROP screening in pre‐mature infants Study funding source: Project of Nanjing Special Fund for Health Science and Technology Development Author COI statement: absence of any commercial or financial relationships

Talebi 2022.

Methods	Study design: between‐group Total study duration: not found
Participants	Total number: 60 Setting: hospital Diagnostic criteria: term infants with a gestational age of 42 to 37 weeks Gestational age: not found Sex "... did not significantly differ in terms of sex" Country: Iran
Interventions	Total number of intervention groups: 3 Control group description: the infants received no (non) pharmacological interventions during the procedure Total number of participants randomised to each of the groups: 15 Specific intervention:  Swaddling group: the neonate was laid on a triangular cloth on a flat surface with no clothes and only a diaper. At first, one side of the sheet, then the bottom, and finally, the other side was folded up over the neonate. In this method, called Frog Flexible, the neonate is able to move the pelvic joints effortlessly, and the arms are bent and placed along the line under the chin, similar to the neonate’s position in the mother’s uterine. The infant’s hand was removed from the swaddle after 10 minutes of swaddling. Sucrose group: sucrose was administered as follows: 0.2 mL/kg of standard 24% sucrose was dumped into the infant’s mouth with a 1 mL needleless syringe 2 minutes before venous blood sampling.  Swaddling + sucrose group: the infants received sucrose and swaddle simultaneously.
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: recorded 2 minutes before, during, and 2 minutes after blood sampling (the observation time for recording the PIPP score was 30 seconds) Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse events: not reported
Notes	Key conclusions of study authors: pain severity during and after venous blood sampling was lower in the swaddle‐sucrose group than in other groups  Study funding source: no funds  Author COI statement: no COI

Usta 2021.

Methods	Study design:between‐group Total study duration: 9 months
Participants	Total number: 76 Setting: hospital Diagnostic criteria: preterm, 24 to 37 weeks of gestation Gestational age: Gestational age (weeks): Experimental group: 32.5 ± 2.3 Control: 33.1 ± 2.8 Sex Males: 32 Females: 29 Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: distilled, odourless water was dropped onto the same type of cotton buds as intervention group Total number of participants randomised to each of the groups: Experimental group: 30 Control group: 31 Specific intervention: lavender oil was dropped onto a cotton bud. The cotton bud was then placed near the baby’s nostrils. The lavender oil was sniffed by the baby 3 minutes before the heel stick procedure, during the procedure, and 30 seconds after the procedure.
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP‐R Time points: 3 minutes before the intervention, during the sampling, and 3 minutes after the procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse events: not found
Notes	Key conclusions of study authors: inhalation of lavender scent is effective in pain control in premature infants Study funding source: no funding Author COI statement: no COI

Yavaş 2021.

Methods	Study design: between‐group Total study duration: 4 months
Participants	Total number: 128 Setting: hospital Diagnostic criteria: between the 37th and 42nd gestational weeks Gestational age: 39.0 (38.0 to 40.0) Sex Males: 61 Females: 67 Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: no intervention before procedure; mother held newborn during heel lance  Total number of participants randomised to each of the groups: 64 Specific intervention: a foot massage was given by mothers with baby oil for 3 minutes before procedure. During the heel lance, the mother was asked to take her newborn in her arms.
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: before during and after Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse events: not found
Notes	Key conclusions of study authors: foot massage performed by the mother reduced the pain level of babies, increased their comfort levels, and decreased their distress levels Study funding source: no funds Author COI statement: no COI

Yu 2022.

Methods	Study design: between‐group Total study duration: 4 months
Participants	Total number: 64 Setting: hospital Diagnostic criteria: premature Gestational age: Length of gestation (weeks) 32.6 ± 5.4 Sex Males: 31 Females: 33 Country: Taiwan
Interventions	Total number of intervention groups: 1 Control group description: infants received containment and non‐nutritive suckling after the procedure if they cried for > 1 minute Total number of participants randomised to each of the groups: 32 Specific intervention: a recording of the maternal voice was acquired in the hospital's postnatal unit by the third postnatal day. On the night of the fourth postnatal day, the prerecorded maternal voice was played for the intervention group infants starting at 3 minutes before a heel stick procedure and lasting until the procedure completion, 13 minutes in total. This was repeated for 3 consecutive days. Infants received containment and non‐nutritive suckling after the procedure if they cried for > 1 minute.
Outcomes	Pain outcomes, scale limits, and time points when measured NIPS Time points: 3 minutes before a heel stick procedure and lasting until the procedure completion, 13 minutes in total Upper limit: 7 Lower limit: 0 High score = more pain Continuous Adverse events: no adverse reactions were noted
Notes	Key conclusions of study authors: the maternal voice intervention slowed the heart rate and alleviated the pain response of the hospitalised premature infants Study funding source: Research Support Scheme of the Chang Gung Memorial Hospital Author COI statement: no COI

Özdemir 2022.

Methods	Study design: between‐group Total study duration: not found
Participants	Total number: 80 Setting: NICU of a research hospital Diagnostic criteria: neonates with hyperbilirubinaemia who were being treated at the NICU; a gestational age of 37 weeks or longer Gestational age: not found   Sex Males: 41 Females: 39 Country: Turkey
Interventions	Total number of intervention groups: 3 Control group description: neonates exposed to routine venipuncture procedure Total number of participants randomised to each of the groups: 20 Specific intervention:  Breast milk smell group: neonates made to smell 2 mL of breast milk dropped on a sterile sponge during the procedure Prone position: neonates moved to the prone position during the procedure Pacifier/dummy group: neonates given a false pacifier/dummy made from sterile gloves
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP‐R Time points: 5 minutes before, during, and after the 5‐minute procedure Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse events: not found any
Notes	Key conclusions of study authors: breast milk smell, prone position, and giving a pacifier made of sterile gloves are effective in reducing the pain and stress of newborns during the venipuncture procedure Study funding source: declared that this study has received no financial support Author COI statement: declared that there were no conflicts of interest

Özkan 2022.

Methods	Study design: between‐group Total study duration: 5 months
Participants	Total number: 60 Setting: hospital Diagnostic criteria: premature  Gestational age: Gestation age (weeks) control: 29.7 ± 1.9 intervention: 29.3 ± 1.5 Sex Males: 32 Females: 28 Country: Turkey
Interventions	Total number of intervention groups: 1 Control group description: no intervention given  Total number of participants randomised to each of the groups: 30 Specific intervention: non‐nutritive sucking with a gloved finger
Outcomes	Pain outcomes, scale limits, and time points when measured PIPP Time points: 60 seconds before the eye examination, 30 seconds during the eye examination, and 30 seconds after  Upper limit: 21 Lower limit: 0 High score = more pain Continuous Adverse events: not found any
Notes	Key conclusions of study authors: pain scores were significantly lower in the intervention group during and after the retinopathy of prematurity eye examinations compared to the control group Study funding source: no specific grant Author COI statement: none

ABSS: Anderson Behavioural State Scale; BPN: Bernese Pain Scale for Neonates; BNBAS: Brazelton Neonatal Behavioral Assessment Scale; COI: conflicts of interest; FLACC: Face Legs Arms Cry Consolability Scale; FT: facilitated tucking; GA: gestational age; GHT: gentle human touch; GT: gentle touch; HR: heart rate; N/A: not available; NFCS: Neonatal Facial Coding System; NIAPAS: Neonatal Infant Acute Pain Assessment Scale; NICU: neonatal intensive care unit; NIPS: Neonatal Infant Pain Scale; NNS: non‐nutritive sucking; N‐PASS: Neonatal Pain, Agitation, and Sedation Scale; PIPP: Premature Infant Pain Profile; ROP: retinopathy of prematurity; SpO2: oxygen saturation

Differences between protocol and review

2011 Review

1. Title was changed from 'Nonpharmacological interventions for needle‐related procedural pain and postoperative pain in neonates and infants', which was the title specified in the protocol, to 'Non‐pharmacological management of infant and young child procedural pain'.

2. We changed our approach to categorising age groups from what was stated in the protocol. We now have three age groups (preterm born neonates, term neonates up to one month old, and older infants from one month up to age three) instead of the four we proposed in the protocol (preterm neonates, term neonates to eight months, eight to 18 months, 18 to 36 months). We made this change since there were not enough studies in older infants to split this group into three. Since this is still a limitation in the literature, we addressed the need for more trials in healthy older infants in the Implications for research section (point #5). 

3. In the protocol, we proposed conducting each analysis separately for each pain measure. However, we made the decision not to analyse studies by this fourth dimension because: 1) there were not enough data on each pain measure across interventions, ages, and pain responses; 2) there is too much variability in outcomes measures (20 pain measures included in this review); and 3) due to the already complex nature of a treatment x age x pain response analysis (the approach we used in this review). Together, these issues do not substantiate further splitting the analyses by a fourth dimension (outcome measure). 

2015 update

1. Only behavioural measures were analysed. Due to significant heterogeneity and missing data, physiological measures were not analysed for our review.

2. The standardised mean difference was analysed instead of the mean difference due to the fact that the included studies measure the same construct but with different scales ‐ refer to Chapter 10 in the Cochrane Handbook for Systematic Reviews of Interventions 10.5.1 (Chapter 10) (Deeks JJ, Higgins JPT, Altman DG (editors). Chapter 10: Analysing data and undertaking meta‐analyses. In: Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (editors). Cochrane Handbook for Systematic Reviews of Interventions version 6.2 (updated February 2021). Cochrane, 2021. Available from www.training.cochrane.org/handbook.).

3. Upon recommendation from Cochrane PaPaS, in the previous update we amended our methods to use a simplified measure of risk of bias based on the Cochrane Handbook for Systematic Reviews of Interventions 5.1.0 (Chapter 8; Table 8.5a) (Higgins 2011a) instead of the Quality of Study Design and Methods Scale (Yates 2005).

2022 update

1. We added the following text to the review to meet current standards:

   1. 'Assessment of risk of bias in included studies' section: "We calculated risk of bias using domains from Chapter 8 in the Cochrane Handbook for Systematic Reviews of Interventions 5.1.0 (Chapter 8; Table 8.5a) (Higgins 2011a)". 

   2. 'Data extraction and management' section: "For data extraction and management, we collated multiple reports of the same study, so that each study rather than each report was the unit of interest in the review. We collected characteristics of the included studies in sufficient detail to populate a table of 'Characteristics of included studies'."

2. We changed the data extraction process since the protocol was published. Specifically, we switched to conducting data extraction using Covidence, a platform for managing systematic reviews, as opposed to extracting into electronic data extraction forms. The review authors in charge of data extraction have changed since the protocol was published due to trainees graduating the lab. We calculated inter‐rater reliability on 20% (as opposed to 30% as proposed in the protocol) of studies due to the large number of studies that we screen. 

3. We changed our approach to calculating the standardised mean difference (SMD) and the standard error of the mean difference (SE) for analysed between‐group and cross‐over trials, consistent with the procedures outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021). We changed the text in the Measures of treatment effect section to illustrate these changes. 

Contributions of authors

• One review author (RPR) created the search strategy under the guidance of four Cochrane‐affiliated librarians (Elizabeth Ulryk, Sylvia Bickley, Caroline Struthers, and Joanne Abbott). 

• One review author contributed to the modification of the original strategies and protocol (RPR). 

• For the update 11 review authors (OB, IS, CC, HG, SB, MD, NR, SAK, KT, DL, LU) independently screened titles and abstracts of studies from the literature searches for inclusion. 

• For the update, 10 review authors performed the extractions (OB, IS, CC, HG, SB, MD, NR, SAK, KT, DL). 

• The articles for the update were located and obtained by OB. 

• For the update three review authors reviewed the extractions and came to consensus (RPR, OB, IS). 

• All review authors for the update were involved in reviewing the manuscript (RPR, OB, IS, CC, HG, SB, MD, NR, SAK, KT, DL, BS, LU). 

• None of the review authors were responsible for assessing or extracting data from articles on which they were authors.

• Dr. Rebecca Pillai Riddell is responsible for the update of this review.

Sources of support

Internal sources

• Lillian Wright Foundation for Maternal‐Child Health, York University, Canada

  Lillian Wright Foundation for Maternal‐Child Health has provided funds to: Ms Oana Bucsea, Ms Hannah Gennis, Ms Shaylea Badovinac, Ms Miranda DiLorenzo, Ms Nicole Racine.  

• York University, Canada

  Operating funds for this project were provided by funds from York University to Dr Pillai Riddell with salary support through the York Research Chairs Program.

External sources

• Canadian Institutes for Health Research, Canada

  Provided operating funds and salary support to Dr Pillai Riddell (Post‐doctoral Fellowship, New Investigator Award), doctoral fellowships to Dr Nicole Racine, Ms Shaylea Badovinac, Ms Miranda DiLorenzo, Ms Kara Turcotte, and Ms Oana Bucsea, a Master's‐level fellowship to Ms Hannah Gennis, as well as graduate support to Dr Sara Ahola Kohut.

• Ontario Graduate Scholarship, Canada

  Provided support to Ms Oana Bucsea, Ms Ilana Shiff, Ms Hannah Gennis, Ms Shaylea Badovinac, Ms Miranda DiLorenzo, Ms Nicole Racine, and Dr Sara Ahola Kohut.

• Social Sciences and Humanities Research Council, Canada

  Provided support to Ms Oana Bucsea, Ms Shaylea Badovinac, Ms Miranda DiLorenzo, Ms Hannah Gennis, Ms Diana Lisi, Dr Nicole Racine and Dr Sara Ahola Kohut.

• Pain In Child Health (Canadian Institutes of Health Research) Strategic Training Initiative, Canada

  Provided support to Ms Oana Bucsea, Ms Hannah Gennis, Ms Shaylea Badovinac, Ms Miranda DiLorenzo, Ms Ilana Shiff, Dr Nicole Racine, and Dr Sara Ahola Kohut.

• National Institute for Health Research (NIHR), UK

  Cochrane infrastructure funding to Cochrane Pain, Palliative and Supportive Care (PaPaS)

Declarations of interest

Dr Rebecca Pillai Riddell CPsych: RPR is a professor and clinical psychologist that specialises in pain and mental health. RPR is a co‐author on six included studies (Hillgrove Stuart 2013; Hogan 2014; Pillai Riddell 2018; Taddio 2015; Taddio 2017; Taddio 2017a).

Oana Bucsea, BSc, MA: No conflicts to declare.

Ilana Shiff, EdM, MA: No conflicts to declare.

Dr Cheryl Chow, BSc, MSc, PhD: No conflicts to declare. CC is a Postdoctoral Fellow and her research focuses on elucidating the aetiology, consequences, and development of paediatric anxiety and pain within the medical context.  

Dr Hannah Gennis, PhD, C Psych (Supervised Practice): No conflicts to declare.

Shaylea Badovinac, MA: No conflicts to declare.

Miranda DiLorenzo, MA: No conflicts to declare.

Dr Nicole Racine, PhD: No conflicts to declare. NR is an Assistant Professor and Child Psychologist who manages patients with child mental health difficulties including anxiety, depression, and behavioural difficulties. 

Dr Sara Ahola Kohut, CPsych: No conflicts to declare. SAK is a Clinical and Health Psychologist and manages patients with chronic disease, with a focus on gastrointestinal disorders.

Dr Diana Lisi, PhD: No conflicts to declare. DL is a Clinical Psychologist (Supervised Practice) working in private practice with clients presenting with a variety of mood and anxiety‐related presentations.

Kara Turcotte, MA: No conflicts to declare.

Dr Bonnie Stevens, RN, PhD: BS receives annual royalties for the Oxford Textbook of Pediatric Pain. BS is a co‐author on three included studies (Ipp 2004; Johnston 1997; Stevens 1999). 

Dr Lindsay Uman, RPsych: No conflict to declare. LU is a Clinical Psychologist working with youth and families, and treats patients with a variety of mental and physical health issues.

Where review authors were also an author on an included study, they were not involved in the data extraction process.

New search for studies and content updated (conclusions changed)