Predictors of Effective Breastfeeding in Newborns and 6‐Week‐Old Infants: Modifiable Pre‐Feeding, Positioning and Latching‐On Elements

ABSTRACT

Aim

Although breastfeeding assessment elements arise from compelling expert opinion, they have generally not been subject to rigorous evaluation. This project aimed to ascertain which elements, if any, are associated with effective breastfeeding and identify key modifiable findings for clinical practice.

Methods

This observational descriptive study analysed the breastfeeding sessions of 300 newborns in the postpartum unit and 166 infants attending their 6‐week young‐child clinic visit, from January to February 2024 in a regional referral hospital in Masaka, Uganda using the Lactation Assessment & Comprehensive Intervention Tool (LAT). Effective breastfeeding was defined as 1 or 2 sucks to 1 swallow by cervical auscultation, and/or observed rocker jaw motion. Associations were analysed using Pearson's Chi‐square test.

Results

Nine elements were significantly related to effective breastfeeding: two pre‐feeding behaviours (infant state, feeding cues); four positioning behaviours (tummy‐to‐mummy, arms around breast, breast not held or shaped, head free of restrictions) and three latching‐on behaviours (gape, head tilt, lower lip reaches first). All but one, head free from restrictions, were also significant for 6‐week‐old babies.

Conclusion

Key findings for clinical practice were identified including evidence for modifiable pre‐feeding, positioning and latching‐on behaviours, which were significantly associated with effective breastfeeding.

Summary.

• Although breastfeeding assessment elements arise from a body of compelling expert opinions, they have generally not been subject to rigorous evaluation.

• The study analysed breastfeeding sessions of 300 postpartum newborns and 166 infants attending their 6‐week young‐child clinic visit using the Lactation Assessment & Comprehensive Intervention Tool (LAT).

• Evidence was found that specific modifiable pre‐feeding, positioning and latching‐on behaviours were significantly associated with effective breastfeeding.

Abbreviations

B‐R‐E‐A‐S‐T

WHO/UNICEF breastfeeding observation form

IBT

ineffective breastfeeding technique

LAT

Lactation Assessment & Comprehensive Intervention Tool

REM

Rapid Eye Movement

SPSS

Statistical Package for the Social Sciences

UNCST

Uganda National Council for Science and Technology

UNICEF‐

United Nations International Children's Emergency Fund

WHO

World Health Organisation

YCC

young‐child clinic

1. Introduction

Despite robust evidence which supports the value of breastfeeding in achieving optimal health for mothers and babies [1], globally, only about 44% of infants are being fed human milk exclusively for the first 6 months after birth, far from the World Health Organisation's (WHO) 2030 target of 70% [1, 2]. Painful breastfeeding is usually among the top reasons mothers offer for early breastfeeding cessation [3, 4]. Studies indicate that this is the result of ineffective breastfeeding technique, which is a prevalent finding in populations around the world [5, 6, 7]. For the mother, specific concerns include mastitis, breast pain, engorgement, inadequate milk production and early cessation of breastfeeding [7]. Paediatric outcomes related to ineffective breastfeeding technique include diarrhoea and acute respiratory infections [8].

Newborns have an innate drive to seek the breast, latch‐on and suckle in the first hour or so after birth [9]. However, for mothers, breastfeeding is a learned behaviour: a modifiable, practiced skill [10]. According to the Merriam‐Webster online dictionary, ‘modify’ means to ‘make basic or fundamental changes, to give a new orientation to or to serve a new end’ [11]. The Cambridge Dictionary defines ‘modify’ as ‘to change something such as a plan, opinion, law or way of behavior slightly, usually to improve it or make it more acceptable’ [12]. When to feed the baby, how to position the baby and oneself and how to move the baby to the breast to latch are all modifiable behaviours [13, 14, 15]. Mothers benefit from individual education and support [13, 16, 17, 18].

For lactation and health care providers, assessment of a breastfeeding session is the initial task in determining precisely what information and support would be helpful for an individual mother. Observation, assessment and documentation of breastfeeding sessions by members of the health care team are a routine component of hospital care in the early hours and days postpartum with the goal of preventing breastfeeding problems and ineffective feeding. Health care team members' competency in evaluating a breastfeeding session is expected according to the Baby‐Friendly Hospital Initiative guidelines [19].

Since the 1970's resurgence of breastfeeding in the United States and other countries, books [20, 21, 22], clinical updates [23], protocols [24] and organisational guidelines [25] written by breastfeeding and lactation experts have described the signs of optimal breastfeeding behaviour without evidentiary support beyond their own experience. The Academy of Breastfeeding Medicine Protocol #26: Persistent Pain with Breastfeeding, for example, acknowledges that most of the management strategies for nipple pain are based on expert opinion.

Although united by a desire to mitigate painful breastfeeding, improve infant outcomes and predict which dyads will need additional breastfeeding support, expert opinion supplies neither consensus nor standardisation in what is expected in breastfeeding behaviour (according to three review papers [26, 27, 28]) beyond the embracive parameters: rooting, positioning, latching, sucking and swallowing.

Assessment tools have been developed to organise and document the breastfeeding session and provide continuity of care in the clinical setting. Tools facilitate the assessment approach by providing elements for the observer to note, usually including those related to positioning and attachment. Although the breastfeeding assessment elements in these tools arise from a body of compelling expert opinions, they have generally not been subject to rigorous evaluation.

In addition, most tools used in research studies aggregate several assessment items into a single category to simplify scoring. For example, a single assessment score may be given for the aggregate of items: ‘Baby well supported, Tucked against mother's body; Lying on side/neck not twisted; nose to nipple; Mother confident handling baby’ [29]. When tools that aggregate the score are used, little is known about individual modifiable breastfeeding elements relative to effective and ineffective feeding techniques that would guide clinical practice.

Two tools do not aggregate: B‐R‐E‐A‐S‐T (WHO/UNICEF breastfeeding observation tool) which has not been validated [28], although it is widely used and the Lactation Assessment & Comprehensive Intervention Tool (LAT) which has been validated and found to be reliable [30]. The LAT allows for the assessment and documentation of 11 specific pre‐feeding, positioning and latching‐on behaviours, as well as two indicators for effective suckling [30]. This provides an opportunity to not only assess a breastfeeding, but also to guide a corrective intervention. The current study focused on the modifiable aspects of the breastfeeding assessment.

Although breastfeeding assessment elements arise from a body of compelling expert opinion, they have generally not been subject to rigorous evaluation. The aim of the current study was to ascertain which modifiable pre‐feeding, positioning and latching‐on elements of breastfeeding, if any, are associated with effective breastfeeding in order to identify key findings for clinical practice.

2. Method

2.1. Study Design and Population

The breastfeeding observations, which provided data for this observational descriptive study, were collected at Masaka Regional Referral Hospital in south‐central Uganda. It is a public hospital funded by the Uganda Ministry of Health; general care in the hospital is free. The hospital reports more than 5500 vaginal births and more than 3000 births by caesarean section annually.

A power analysis was conducted using G*Power (version 3.1.9.7) for a Chi‐Square Goodness‐of‐Fit Test, applying a medium effect size (w = 0.30) to account for the sample. The analysis indicated that a sample size of 145 would provide 95% power to detect significant differences at an alpha level of 0.05.

The study population was a convenience sample of two cohorts: 300 newborns born to women who arrived to give birth at Masaka Hospital between January 3 and 10, 2024 and from February 21–27, 2024 and 166 babies who were brought to the young‐child clinic (YCC) for their 6‐week‐old vaccinations during the same periods in January and February of 2024.

Each mother who came to the hospital in labour, or who was scheduled for caesarean section, was approached to give consent for their baby to be assessed while breastfeeding before discharge. Mothers who brought their babies for their 6‐week visit to the YCC were approached to provide consent at the time of the visit. Mothers were asked to breastfeed in their usual way.

The study was described to each mother, and she had the option to consent to the study or not. Healthcare providers at the facility had the option to deselect consented participants for health reasons. Names and identifying information were removed from the data collection material to safeguard private, protected health information. A unique code was generated for each dyad and recorded on the research copy of demographic data, field notes and on the LAT. For hospital dyads, demographic information was retrieved from the hospital's paper charting system. Demographic information at the 6‐week visit to YCC was obtained by research assistants in the mother's preferred language. Gestational age for the dyad in the newborn group was estimated via fundal height measurement upon admission. Mothers in the 6‐week visit to YCC group were asked about their baby's gestational age at birth, and this was recorded. Each dyad's individual code was entered on a paper copy of the LAT which was used to document the findings of the breastfeeding assessment. All materials were kept in a locked room during the study and in password‐protected computers after the data was entered.

2.2. The Tool

The LAT includes 11 pre‐feeding, positioning and during latching‐on items, as well as two items that assess suckling effectiveness. It also includes other items related to assessing the attachment during feeding and the condition of the dyad after the feeding, which are not included in this analysis.

The pre‐feeding items include whether the infant is held skin‐to‐skin immediately prior to/during this current feeding; the infant's state (deep sleep, light sleep, quiet/active alert or crying) and whether and which feeding cues were observed. Feeding cue item choices include rooting, hand‐to‐mouth actions, mouthing/sucking motions, Rapid Eye Movement (REM) and body movements. Also included is the choice of no cues observed prior to feeding. Note that crying is assessed as a state and not part of the category of feeding cues.

The group of items assessed in the positioning section of the LAT includes whether the baby's body is turned towards the mother and if the head, shoulders and hips are in alignment (‘tummy to mummy’). The infant's arm positions are also assessed; optimally, the arms should be around the breast, not between the mother's body and the baby. The infant's head should be free of restrictions, and the breast neither held nor shaped.

The latching‐on behaviour items include whether the nose is opposite the nipple to start, whether there is a gape response during the latching‐on behaviour, whether the head is free to tilt back and whether the bottom lip reaches the breast first.

To determine the effectiveness of the feeding, swallowing is confirmed by cervical auscultation with a neonatal stethoscope. Bursts of 1:1 and/or 2:1 suck‐to‐swallow ratios are associated with adequate milk transfer as is the rocker jaw motion (compared to the piston motion). Suckling and milk transfer are associated with breastfeeding duration [31]. Validity [32, 33] and reliability [30] have been established for the LAT. To ensure interrater reliability, the three assessors jointly observed and scored the first 20 dyads, discussing any discrepancies until a consensus was reached. The level of agreement was high and any subsequent discrepancies that arose during data collection were resolved through discussion and consensus.

2.3. Data Analysis

During the breastfeeding observations, the assessors documented the assessment on a paper copy of the tool, which was labelled with the dyad's unique number. The relationship of each of the pre‐feeding, positioning and latching‐on items to the suck‐to‐swallow ratio and/or rocker/piston motion of the jaw was analysed using Statistical Package for the Social Sciences (SPSS) software version 26 using Pearson's Chi square.

2.4. Ethics

Ethics approval was obtained from the Makerere University School of Health Sciences Research Ethics Committee in Kampala, Uganda (MAKSHSREC‐2023‐558) and the Uganda National Council for Science and Technology (UNCST) registration number HS3183ES. Masaka Regional Referral Hospital also provided Administrative Permission/Clearance. All mothers provided written informed consent for themselves and their infants to participate in the study in their own language and were free to discontinue participation at any time without affecting their care.

3. Results

3.1. Demographic Results

A comparison of the cohort groups' demographic data (Table 1) indicates no significant differences in the two groups, with one exception. The gestational age of the infants was significantly different between the two groups (p < 0.001). This may be because of the way the information was collected. For the newborn group, gestational age was obtained from the hospital record as determined by the midwife's fundal height measurement during the hospital admission examination. For the cohort of 6‐week‐olds, gestational age was obtained during the consent process from the mother. It should be noted that neither of these two methods of assessment are precise in estimating gestational age [34, 35]. None of the dyads were deselected.

TABLE 1.

Demographic table for postpartum and 6 week dyads.

	Postpartum; N = 300	6 weeks postpartum; N = 166	Significance
Mother's age (years)	26.30 ± 5.91	26.35 ± 5.90	0.228
Number of previous pregnancies	2.70 ± 1.95	2.37 ± 1.68	0.109
Number of previous live births	1.53 ± 1.76	1.27 ± 1.60	0.272
Gestational age (weeks)	38.11 ± 1.19	38.87 ± 2.01	< 0.001
Newborn weight (kg)	3.159 ± 0.51	3.13 ± 0.61	0.071
Gender of baby (male/female)	133/145	74/83	0.770
Mode of birth (vaginal/caesarean)	252/48	136/30	0.256

3.2. Assessment Findings

Table 2 shows the findings for newborn dyads, and Table 3 shows the findings for 6‐week‐old dyads. The infant's pre‐feeding behaviours, including the infant's state and feeding cues, were positively associated with optimal feeding. No babies in either cohort were assessed to have had skin‐to‐skin contact immediately prior to the breastfeeding assessment or nose opposite the nipple to start, so these items on the tool were not included in the analysis.

TABLE 2.

Breastfeeding behaviours and effective feeding outcomes for newborn dyads.

	Suck to swallow ratio of 1:1 or 2:1	Full rocker jaw motion while feeding
Pre‐feeding behaviours
Skin‐to‐skin prior to feeding	Not observed	Not observed
State (active alert)	χ 2 (3, N = 226) = 11.17, p = 0.011	χ 2 (3, N = 205) = 6.24, p = 0.100
Feeding cues are observed	χ 2 (1, N = 219) = 23.64, p < 0.001	χ 2 (1, N = 200) = 8.781, p = 0.003
Positioning behaviours
Tummy to mummy	χ 2 (1, N = 251) = 31.69, p < 0.001	χ 2 (1, N = 231) = 23.75, p < 0.001
Arms around breast	χ 2 (1, N = 249) = 8.81, p = 0.003	χ 2 (1, N = 229) = 2.23, p = 0.135
Baby's head free of restrictions	χ 2 (1, N = 251) = 7.20, p = 0.007	χ 2 (1, N = 231) = 4.48, p = 0.034
Breast not held or shaped	χ 2 (1, N = 248) = 17.72, p < 0.001	χ 2 (1, N = 229) = 13.09, p < 0.001
Latching‐on behaviours
Nose opposite nipple	Not observed	Not observed
Gape observed	χ 2 (1, N = 221) = 26.43, p < 0.001	χ 2 (1, N = 204) = 6.82, p = 0.009
Head tilt during latch‐on	χ 2 (1, N = 222) = 19.35, p < 0.001	χ 2 (1, N = 204) = 29.96, p < 0.001
Lower lip reaches breast before upper lip	χ 2 (1, N = 216) = 7.78, p = 0.005	χ 2 (1, N = 200) = 24.38, p < 0.001

TABLE 3.

Breastfeeding behaviours and effective feeding outcomes for 6‐week‐old dyads.

	Suck to swallow ratio of 1:1 or 2:1	Full rocker jaw motion while feeding	Did at least one aspect of effective feeding show significance?
Pre‐feeding behaviours
Skin‐to‐skin prior to feeding	Not observed	Not observed
State (Active Alert)	χ 2 (3, N = 152) = 9.64, p = 0.022	χ 2 (3, N = 148) = 1.88, p = 0.597	Yes
Feeding Cues are observed	χ 2 (1, N = 151) = 3.859, p = 0.049	χ 2 (1, N = 147) = 1.74, p = 0.187	Yes
Positioning behaviours
Tummy to mummy	χ 2 (1, N = 155) = 4.44, p = 0.035	χ 2 (1, N = 151) = 9.26, p = 0.002	Yes
Arms around breast	χ 2 (1, N = 150) = 3.76, p = 0.053	χ 2 (1, N = 146) = 6.26, p = 0.012	Yes
Baby's head free of restrictions	χ 2 (1, N = 155) = 2.83, p = 0.092	χ 2 (1, N = 151) = 2.32, p = 0.128	No
Breast not held or shaped	χ 2 (1, N = 155) = 6.04, p = 0.014	χ 2 (1, N = 151) = 12.52, p < 0.001	Yes
Latching‐on behaviours
Nose opposite nipple	Not observed	Not observed
Gape observed	χ 2 (1, N = 149) = 16.59, p < 0.001	χ 2 (1, N = 145) = 3.43, p = 0.064	Yes
Head tilt during latch‐on	χ 2 (1, N = 150) = 5.15, p = 0.023	χ 2 (1, N = 146) = 9.94, p = 0.002	Yes
Lower lip reaches breast before upper lip	χ 2 (1, N = 148) = 2.66, p = 0.103	χ 2 (1, N = 146) = 5.95, p = 0.015	Yes

4. Discussion

It is a common issue globally that babies are positioned sub‐optimally at the breast [6], with the prevalence of ineffective breastfeeding technique (IBT) ranging from 30% to 70% [6]. Breastfeeding technique and breastfeeding knowledge are significantly associated with exclusive breastfeeding [36] and an absence of breast problems [37, 38]. Comparison with other studies focusing on specific elements of breastfeeding is not possible because after bibliographic review, we could find no other studies that compared specific breastfeeding elements with effective sucking. Effective breastfeeding is an infant and child survival issue. Ineffective latch and positioning can have dire consequences [5]. Research about the evaluation of correct positioning often does not include the modifiable elements included in this study—state, pre‐feeding behaviours and the activity of attaching [39].

Breastfeeding interaction, or ‘technique’, is commonly characterised as a learned skill for mothers; a combination of three elements: positioning, attachment and sucking [7]. It has been understood that these are the three attributes that contribute to effective feeding. ‘Effective breastfeeding is a function of the proper positioning of mother and infant baby and attachment of child to the mother's breast’ [8] (p. 309). Our findings both confirm and challenge the positioning, attachment and suckling construct. This research confirms the importance of the three elements, but expands the construct to include two more; pre‐feeding behaviours and latching‐on behaviours (which include the action of moving to the breast) were also significantly related to effective feeding.

Brazelton identified infant states to include deep sleep, light sleep, quiet alert, active alert and crying [40]. Quiet and active alert are acknowledged to be the effective states to begin breastfeeding [41, 42], ‘because this is the state when the baby is most attentive and available for task mastery’ [43] (p. 293). However, research has been lacking on the connection between baby state and effective breastfeeding. In the current study, when a nursing session began with the observed baby's state of quiet/active alert, the chances were significant that swallowing would be heard, both for newborn infants in the postpartum unit and 6‐week‐old infants.

Feeding cues are affirmed to be a good indicator that the baby is ready to feed [44, 45]. For preterm infants, cue‐based feeding is associated with decreased time to full oral feeds and less feeding intolerance [46]. In full‐term infants, the use of feeding cues is associated with a healthy relationship between the mother and baby [47]. To our knowledge, this is the first research that found an association between feeding cues and effective sucking. The observation of feeding cues, including REM sleep, mouthing motions and rooting, prior to beginning a breastfeeding session was especially important for newborns.

Weigert and colleagues found a significant relationship between baby's head and trunk not aligned and not exclusive breastfeeding [48]. ‘Tummy to mummy’ in the LAT includes two categories from Davra et al.'s operational definitions: ‘head and body straight’ and ‘infant's body close to mother's body’ [49]. In the Davra et al. study, positioning was not investigated in relation to effective breastfeeding; ‘good’ positioning was correlated with mothers who are in a nuclear family and have low parity [49].

To our knowledge, this is the first research that associated the positioning of tummy to mummy with effective sucking. The alignment of the baby's head, shoulders and hips to the mother's body prior to beginning a breastfeeding session was important for newborns and 6‐week‐old infant's breastfeeding effectiveness.

When a newborn massages the mother's breast during a feeding, the mother's oxytocin levels rise [50], which may impact uterine contraction and milk ejection [51]. The baby's arms around the breast prior to beginning a breastfeeding session were significant for newborns and 6‐week‐old infants' breastfeeding effectiveness.

The baby's head is free of restrictions when the mother's hand is placed on the baby's shoulder blades, and not on the back of the neck. Ingram notes this ‘mum's hand behind shoulders’ as one of their eight elements in a breastfeeding technique [52]. Other times it's discussed as supporting their back and not their neck or head. For example, La Leche League International recommends.

‘pull your baby close by supporting their back (rather than the back of their head)’ [53] Although these positioning techniques are frequently recommended, extensive bibliographic review did not find connections between effective breastfeeding and baby's head free from restrictions. The current study found that baby's head free of restrictions was significantly associated with effective breastfeeding both for newborns and for 6‐week old infants.

Thompson found that the breast being held or shaped for the purpose of directing the nipple towards the baby's mouth was significantly associated with both engorgement and nipple trauma [54]. Altuntaş and Ünsal found no difference between methods of breast holding (c‐hold versus scissor‐hold) for milk intake by the infant. However, they did not include a group without breast holding or shaping [55]. We found that the mother not holding or shaping her breast prior to beginning a breastfeeding session was significant for newborns and 6‐week‐old infants' breastfeeding effectiveness.

Breastfeeding assessment tools may focus on baby's state, observation of feeding cues, positioning, attachment and effective feeding [39, 56]. However, attachment in these tools refers to the assessment of the baby already attached to the breast [39, 56]. The LAT includes the active process of attaching to the breast [30]. This includes the gape, the head tilting back as the baby approaches the breast, and the lower lip reaching the breast first. All are frequently included in breastfeeding education [57, 58].

Ingram and colleagues found that their breastfeeding technique, which included a wide open mouth during latching‐on, was associated with exclusive breastfeeding at 2 and 6 weeks and a decrease in feelings of perceived milk insufficiency [52]. The baby gaping during the attaching process, as the baby comes to the breast, was significant for newborns and 6‐week‐old infants' breastfeeding effectiveness.

A bibliographic review could not connect the common recommendations of the baby's head tilting back during gape or the baby's bottom lip and tongue reaching the breast first to an effective breastfeeding outcome. However, they were significant to effective breastfeeding in this study.

This study has the merit of being the first study to connect specific elements of breastfeeding technique to effective breastfeeding. Future research should explore intervention strategies focusing on these key behaviours and assess their impact on breastfeeding success in diverse settings.

This research also has implications for training and policy. By identifying specific, modifiable behaviours associated with effective breastfeeding, these findings can inform targeted training for healthcare providers and policy efforts to improve breastfeeding support in clinical settings.

5. Conclusion

Breastfeeding assessment elements arise from a body of compelling expert opinions. However, even though ineffective breastfeeding can potentially lead to a multitude of problems for both the mother and the baby, it has generally not been subject to rigorous evaluation. This study analysed breastfeeding sessions of 300 postpartum newborns and 166 infants attending their 6‐week young‐child clinic visit. The findings change the construct of breastfeeding assessment, providing specific criteria for effective intervention. Evidence was found for the impact of modifiable pre‐feeding, positioning and latching‐on behaviours which were significantly associated with effective breastfeeding.

6. Limitations

Mothers were asked to breastfeed in the post‐partum unit when they were available to the researchers, often directly prior to discharge. The babies may not have been in the optimal state for the assessment, and this may have affected some outcomes.

Gestational age was estimated by fundal height for newborns and by maternal recall for 6‐week‐old infants. Both methods are subject to imprecision, which may have introduced some misclassification and affected comparisons between groups.

Pre‐ and post‐weights for milk transfer were not possible in this study. Cervical auscultation was the method of confirming swallowing.

Attachment elements (how the baby is latched to the breast) are not modifiable and so were not included in this analysis.

None of the babies in the study were being held skin‐to‐skin directly prior to feeding, and none of the babies latched with their noses opposite the nipples, so neither of these modifiable elements of the latching process could be assessed or analysed.

For dyads approached during the 6‐week YCC visit, the 6‐week‐old babies of some mother/infant dyads were assessed pre‐vaccination, and some were assessed post‐vaccination. It is unknown if this difference impacted the infant's state. It is possible that the baby fed either more or less effectively after being vaccinated.

Given the number of elements assessed, no formal adjustments for multiple comparisons were made. The findings should therefore be interpreted with caution regarding potential type I error.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.