Basic life support training targeted to family members or carers of those at high-risk of out-of-hospital cardiac arrest: a systematic review

Abstract

Introduction

This updated systematic review examined the impact of targeting basic life support (BLS) training to groups at high risk of out-of-hospital cardiac arrest (OHCA) on patient and educational outcomes.

Methods

Searches of databases (MEDLINE, EMBASE, CINAHL and Cochrane Library) were conducted from June 2014 to November 2024 (PROSPERO CRD42021233811). We identified pre-2014 studies from the 2015 International Liaison Committee on Resuscitation evaluation of this topic. Two reviewers independently screened titles and full-text papers, extracted data and assessed risk of bias (ROB-2 and ROBINS-I). The certainty of the evidence for each outcome was evaluated using GRADE methodology. A narrative synthesis was performed.

Results

A total of 48 studies (17 new non-randomised studies) were found. While there is evidence of subsequent use of skills and improved survival with the intervention, the quality of evidence remains very-low to low, due to the low number of OHCA events during follow-up and significant loss to follow-up. For educational outcomes, the quality of evidence for CPR skills, confidence and willingness to perform CPR was low to moderate, with most studies showing improvements in these outcomes or skills at guideline recommendations following training. Those trained were also likely to train others.

Conclusions

Despite a large number of studies examining BLS training to those at high-risk of OHCA, the quality of evidence remained largely unchanged. However, the evidence continues to demonstrate that targeting those at high-risk improves simulated CPR skills, and confidence and willingness to provide CPR in likely rescuers.

Introduction

Out-of-hospital cardiac arrest (OHCA) is a sudden and distressing event for lay rescuers, particularly family members who are likely to be witnesses.1, 2 Survival depends on implementing the ‘chain of survival’,³ including basic life support (BLS) by rescuers. Every minute BLS is not provided, the probability of survival decreases by 10 %,⁴ but bystanders administering CPR alone is associated with a two to threefold increase in survival.5, 6

Training in BLS is widely available to the public, but is not always taken up due to low awareness, time and costs.⁷ National BLS training rates are reported between 3 % and 81 %.⁸ However, this includes those who have ever undertaken BLS training and rates of recent training are very low.⁷ Targeted BLS training in family members and carers of those at high-risk of OHCA was first recommended by the International Liaison Committee on Resuscitation (ILCOR) in 2015⁹ to ensure those with a higher risk of witnessing a cardiac arrest are trained, and also as a method to increase secondary training (i.e. those who are trained go on to train others) of other members of the community.

The ILCOR reviews, published in part in the annual ILCOR publications,9, 10 and another earlier systematic review¹¹ on this topic in adults, showed strong indications that training would be rescuers improves their BLS knowledge and skills performance, along with increases in willingness to provide BLS without increasing their anxiety levels. However patient outcomes, such as survival, are more difficult to assess due to a very low quality of evidence caused by insufficient number of events and loss to follow-up.9, 10, 11 Since these reviews were conducted there have been further studies, including studies reporting patient outcomes. Therefore, this systematic review aims to update and present in full, the existing evidence on the effectiveness of targeted BLS training strategies to those at high-risk of OHCA.

Methods

Before commencing the literature search, this systematic review was registered in the International Prospective Register of Systematic Reviews (PROSPERO CRD42021233811). We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) 2020 expanded checklist (Supplement 1).¹² As this review used an updated search strategy (Table S7, Supplementary Materials) and included studies from the 2015 ILCOR review⁹ (SC, JF and JB were authors of the 2015 review), we consider this review an updated systematic review. As the timing of reported patient survival outcomes varied or was not reported at all (e.g. reported as survived/died and not survival to hospital discharge), we included all patient outcomes irrespective of timing and report on them as described in the studies.

PICO question and outcomes

The PICO question for this review was defined as: For adults and children at high-risk of OHCA (P), does focussed BLS training of likely rescuers (e.g. family or care-givers) (I), compared with no such training (C), change patient or educational outcomes (O). Although this review was conducted outside the ILCOR Education, Implementation and Teams (EIT) Task Force, the PICO and outcomes were approved by the EIT task force before PROSPERO registration. Outcomes are listed in Table 1.

Table 1.

Systematic review outcomes of interest.

	Outcome	Level of outcome
Patient outcomes	Survival with favourable neurological outcome at discharge and 30 days*	Critical outcome
	Survival to hospital discharge/30 days*
	Return of spontaneous circulation*
	Rates of bystander CPR (subsequent use of skills)
	Rates of AED use (subsequent utilisation of skills)
	Bystander CPR quality during OHCA (any available CPR metrics)
Educational outcomes^	CPR quality and AED competency post training completion and within 12 months of training	Important outcome
	CPR and AED knowledge post training completion and within 12 months of training
	Confidence to perform CPR post training and within 12 months of training
	Willingness to perform CPR post training and within 12 months of training
	Secondary training of others#

CPR, cardiopulmonary resuscitation; AED, automated external defibrillator; OHCA, out-of-hospital cardiac arrest.

^*Any measure and timeframe of neurological outcome and survival were included.

^{^}All measures used were included.

^#Secondary training = participants training others.

Eligibility criteria

We included all studies included in the previous ILCOR review and used the same eligibility criteria. We included studies that reported BLS training to adults (any person over 18 years) where they were the family member, carer or someone with a close association to a person at high-risk of OHCA. The definition of those in a high-risk OHCA population include a person of any age who has cardiac disease, experienced a prior OHCA, a person who uses drugs, or any other illness with high-risk of cardiac arrest. We accepted any method of BLS training, as well as studies comparing different training methods and studies with no control groups. To be included, studies had to report on any of the pre-specified patient or educational outcomes detailed in Table 1. Our primary and critical outcomes relate to OHCA events that occurred in high-risk patients during the period of follow-up after the BLS training was provided.

We included published original research articles of randomised and non-randomised interventional studies. We included observational studies to complement randomised evidence as these studies are often more “real world” using more diverse populations. Case series studies were eligible if the number of cases include was greater than five. Commentary, editorials, review and animal studies were excluded. Studies published in languages other than English were considered if a translator could be sourced.

Search strategy and information sources

We updated the search strategy (Table S7, Supplementary Materials) from the 2015 ILCOR review,⁹ which was generated with assistance from an information specialist. The new search was conducted on 12 November 2024, and included publications from June 23rd 2014. Searches were conducted using the following databases; PubMed (1996-), EMBASE (1966-), CINAHL (1937-) and Cochrane Library. There was no restriction on language (provided an English abstract was available). Reference lists of included studies were also searched.

Study selection

Search strategy results were imported into COVIDENCE online software¹³ to assist with screening. Titles and abstracts were independently screened by at least two authors (ZD, SC) against the inclusion criteria. Disagreements at this stage were discussed with another author (JB). Full text publications were independently appraised by two authors (ZD, SC). Disagreements at this stage were discussed with a third author (JB).

Data extraction

One author (ZD) independently extracted data from each study using a pre-piloted electronic data extraction form. This was checked by a second author (SC) and discrepancies were resolved by discussion. Extracted information for each study included: author(s), publication year, study location and design, sample description and size, type and modality of BLS training and outcomes of interest. As the studies used different controls and assessment tools to measure educational outcomes, data for these outcomes are summarised as improved/unchanged/worsened from baseline or above/at/under the guideline standard (e.g. compression rate with guideline recommendations at the time of the study).

Risk of bias assessment

Risk of bias was assessed for all included studies independently by two authors (ZD, JB). We used the Cochrane Risk of Bias 2.0 (RoB 2.0) tool for randomised controlled trials (RCT) and the Risk of Bias in Non-Randomised studies (ROBINS-I, v19) assessment tool for non-randomised studies14, 15 Following study level assessments, we used the Grades of Recommendation, Assessment, Development and Evaluation (GRADE) approach for assessment of the certainty of evidence at the outcome level.¹⁶

Data synthesis

Similar to the previous reviews, we expected studies to be heterogeneous, thus precluding meta-analysis, therefore we planned to present a narrative analysis of studies using published methods.¹⁷

Results

Study selection

We included 31 studies (4 RCTs2, 18, 19, 20 and 27 non-RCTs21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47) from the previous review with the updated search identifying 17 new non-RCT48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64 studies (Fig. 1). One study from the 2015 review only examined anxiety levels⁶⁵ and was therefore excluded from this update. Five new studies were randomised trials, but for the purpose of this review they are classified as non-RCTs as they compared different methods of BLS training with no control group.49, 50, 54, 58, 61

Fig. 1.

Prisma flow diagram.

Study characteristics

The characteristics of the 48 included studies are detailed in the Supplementary materials (Table S1 and S2. Studies were predominantly conducted in the United States of America (USA, n = 30)2, 18, 19, 30, 54, 22, 23, 24, 25, 26, 27, 28, 33, 34, 35, 36, 37, 39, 40, 41, 42, 45, 46, 47, 48, 49, 50, 51 and Europe (n = 9)20, 29, 32, 38, 43, 44, 59, 60, 62 and were published between 1986 and 2024.

High-risk adult populations were targeted in 28 studies,2, 19, 20, 24, 31, 32, 37, 38, 40, 43, 44, 46, 58, 61, 63, 27, 28, 29, 48, 49, 50, 52, 53, 54, 55, 56 with 20 studies18, 25, 30, 39, 41, 42, 45, 47, 51, 57, 59, 60, 22, 23, 33, 34, 35, 36 focused on high-risk paediatric populations (Table S1 and S2). Adult studies predominantly trained family members of cardiac patients (n = 23),2, 19, 20, 31, 37, 38, 40, 43, 46, 49, 50, 58, 61, 63, 27, 28, 52, 53, 54, 55, 56 three studies targeted family members of cardiac arrest survivors,29, 32, 44 and one in women with a history of drug use.⁴⁸ Paediatric populations varied, with the largest proportion targeting family members of infants in the neonatal intensive care unit (NICU) (n = 8).18, 22, 23, 25, 36, 41, 57, 60

Studies used in-person and self-learning training modalities (video-self instruction, computer, mobile application), in two studies the type of CPR training was not described.30, 48 Training was held in-hospital (inpatients or outpatients) and at home; in 13 studies the training location was not described. 28, 30, 31, 33, 45, 37, 38, 39, 59, 60, 61, 62, 63 Included studies ranged in sample size from 26³⁷ to >1400⁴⁹ participants for non-randomised studies, and 65¹⁹ to 7001²⁰ in the RCTs.

Seventeen studies (including 3 RCTs) included follow-up to examine whether CPR skills were used for OHCA events and subsequent patient outcomes.23, 26, 27, 29, 30, 33, 37, 38, 42, 43, 48, 57, 59, 19, 20 This follow-up was typically obtained by contact with participants. Most of the studies (n = 24, 1 RCT) examined the ability to perform CPR skills following training (either on completion of training or in the following 12 months).2, 21, 22, 39, 40, 45, 47, 58, 60, 61, 25, 26, 35, 36, 37, 49, 50, 51, 52, 53, 54, 55

Risk of bias within studies

The risk of bias assessment is presented in Supplementary Tables S3 and S4. Two of the RCTs were deemed to have a low risk of bias overall,2, 20 with the remaining two RCTs assessed as critical due to randomisation, missing data and outcome measurement.18, 19 Most non-RCTs (n = 10)23, 26, 27, 29, 30, 33, 42, 43, 48, 57 were at serious or critical risk of bias for patient outcomes. For educational outcomes, nine non-RCTs24, 25, 28, 31, 39, 40, 44, 55, 60 were assessed as having low risk of bias. The remainder (n = 27) ranged from moderate to critical risk of bias26, 29, 32, 22, 23, 34, 35, 36, 37, 41, 42, 56, 58, 45, 46, 47, 49, 50, 51, 52, 53, 54, 61, 62, 63, 64, most (n = 19) frequently due to high rates of missing data.22, 26, 29, 35, 37, 41, 42, 49, 50, 56, 61, 62, 46, 47, 52, 53, 54

Outcomes

The variation between studies for methods of training and outcome measurement precluded any meta-analysis. The overall certainty of evidence for critical patient outcomes was very-low, downgraded for serious risk of bias and imprecision (Table S5). The certainty of evidence for important educational outcomes was predominantly low, most commonly downgraded for serious risk of bias (Table S6). Only two outcomes, CPR skills and AED competency at training completion and willingness, were considered to be at moderate certainty of evidence, upgraded for consistency. No studies were found where quality of CPR was measured during an OHCA.

Many studies (n = 9) also did not report outcomes in the pre-specified timeframes (e.g. hospital discharge), with some studies reporting outcomes without definition, use of a formal measure (e.g. cerebral performance category scale) or timeframe (e.g. simply reported as “successful resuscitation”, “recovery”, “good neurological function”) (Table 2). 18, 19, 26, 33, 37, 38, 42, 59, 62 Given the low number of OHCA events, we accepted any definition and timeframe of neurological outcome and survival used by the authors.

Table 2.

Patient outcomes after BLS training to family members and carers.

Author	High-risk populations	Outcome (follow-up period)	Result
RCTs
Bardy 200820	Adult cardiac (n = 7011)	Use of AED and survival to hospital discharge (median 37.3 months follow up)	100 % follow-up 160 OHCAs (82 intervention), 38 survived 48 h (19 intervention) 32 study AED uses (29 patients were unresponsive), 12 received defibrillation, 4 survived to hospital discharge
Dracup 198618	Adult cardiac (n = 65)	Use of CPR (within 6 months) and survival	100 % follow-up 4 OHCAs (2 intervention), none received CPR, all died
Dracup 200017	Infants in NICU (n = 462)	OHCA in the home, CPR provided, successful resuscitation (within 12 months)	58 % follow-up 13 OHCAs in the home (0 control group), all successfully resuscitated
Non-RCTs
Ataiants 202048	Women using illicit drugs witnessing an overdose (n = 220)	Use of BLS skills when witnessing an overdose (in the last 12 months)	n/a follow-up Increased naloxone use in trained women (20 % vs. 4 %, p = 001) No difference in use of CPR or rescue breaths
Dracup 199426	Adults cardiac (number of patients not provided)	Use of CPR in trained family members (mean 21 months after training)	72 % follow-up 1 OHCA among high-risk patients, no CPR (trained person not present), did not survive 4 instances of CPR use by trainees, 3 survivors
Dracup 199823	Convenience sample of parent of infants in Dracup 1998 RCT (number of infants not provided)	CPR use on infants (within 6 months) and patient outcome	94 % follow-up 7 used CPR on infants (respiratory arrests), all successfully resuscitated
Eisenberg 198927	Adult ventricular fibrillation cardiac arrest survivors (n = 97)	AED use and survival to hospital and hospital discharge (up to 57 months)	100 % follow-up 14 OHCAs (10 in AED group), AED used in 6 (2 were in VF) 7/14 survived to hospital (5 AED group), 2/14 survived to discharge (1 AED group),
Haugk 198929	Cardiac arrest survivors (n = 88)	AED use (within 12 months)	94 % follow-up AED applied 1/88, normal rhythm
Higgins 198930	Paediatric cardiac hospitals (n = 41)	Parental bystander CPR and survival for OHCAs (over 10 years)	83 % follow-up 65 OHCAs: in the 41 OHCA from hospitals teaching CPR to parents, 68 % received CPR from parents and 31 % survived (46 % of those who received CPR survived); No CPR or survival was seen in the 24 OHCAs from non-CPR teaching hospitals
Jisha 202257	Preterm babies (n = 126)	Parental bystander CPR and recovery at home (unknown follow-up period)	47 % follow-up 5 babies received CPR from parents, 4 recovered quickly at home
Knight 201333	Children high risk (n = 117)	Use of CPR* by trainees (within 6 months of training) and survival*	52 % follow-up 5/61 trainees performed CPR*, 4 survived with good neurological outcome*
McDaniel 198837	Adult cardiac patients (n = 16)	AED use (within 3 months) and survival	100 % follow-up 1 OHCA (trainee not present), died
McLauchlan 199238	Adult cardiac patients (n = 27)	BLS use (within 2 years)	77 % follow-up 1 OHCA (CPR not performed due to bystander disability), died
McLeod 201759	Children high risk cardiac (n = 44)	AED use and patient outcomes (mean follow-up = 75.5 months)	100 % follow-up AED used 4 children, 3 OHCA, 2 defibrillated, 2 survived with good neurological outcome
Pierick 201242	Children premature or cardiac disease (n = 311)	CPR use (within 12 months) and patient outcomes	41 % follow-up 8/127 events, 7/127 received CPR by a parent, 6/8 OHCAs survived all had good or stable neurologic status
Sanna 200643	Adult cardiac patients (n = 89)	CPR/AED use (within 12 months) and patient outcomes	Follow-up rate not reported No events
Tomatis-Souverbielle 201962	Children acute life-threatening events, apnoea episodes, and choking (n = 62)	CPR use (within 3 months) and patient outcomes	Unknown follow-up 4/62 OHCAs, 4/62 CPR knowledge and skills were applied, all recovered well

RCT, randomised controlled trial; AED, automated external defibrillator; OHCA, out of hospital cardiac arrest; BLS, basic life support; CPR, cardiopulmonary resuscitation.

^*Not pre-specified outcomes and unclear if outcomes are for high-risk patients.

Subsequent use of BLS skills and patient outcomes

Seventeen studies18, 19, 20, 23, 26, 27, 29, 30, 33, 37, 38, 42, 43, 48, 57, 59, 62 (including 3 RCTs18, 19, 20), reported at least one critical outcome, with periods of follow-up ranging from 3 months³⁷ to 10 years³⁰ and most studies relied on self-reported outcomes (Table 2). Both the adult (n = 919, 20, 26, 27, 29, 37, 38, 43, 48) and paediatric (n = 818, 23, 30, 33, 42, 57, 59, 62) studies often reported significant loss to follow-up with few subsequent OHCA events to making the effect of the interventions on patient outcomes unclear (Table 2).

Among the studies examining patient outcomes, there were two large RCTs.18, 20 The Home Automated External Defibrillator Trial (HAT), randomised 7001 adult patients with acute myocardial infarction to have an AED with BLS training compared to CPR training alone.²⁰ In this study, with 100 % follow-up, 160 OHCAs occurred over a median follow-up of 37 months, but only 58 (36 %) arrests were witnessed by trained family members. It is unclear how many received CPR, but 29 (50 %) patients had an AED applied. Notably, in this study, there were seven instances of study AEDs being used for individuals not included in the study (e.g. neighbours). The largest paediatric RCT, which trained parents of 462 infants discharged from a neonatal intensive care unit, compared three methods of CPR training in addition to a control group with no training.¹⁸ At one year, only 58 % were followed up, with parents of these infants reporting 13 OHCA events in the home. All of these children were successfully resuscitated (not defined) and all had received CPR training, with no events reported in the control group.

CPR and AED quality, competency and knowledge

For the important educational outcomes of CPR quality and AED competency at training completion, 19 studies (1 RCT,² 18 non-RCTs21, 22, 25, 39, 40, 45, 47, 58, 60, 36, 37, 51, 52, 53, 54, 55) were found, 8 of which were new in this updated review51, 52, 53, 54, 55, 58, 60, 61. For non-RCTs the certainty of evidence was upgraded from very-low to low for consistency in findings and RCT certainty of evidence remained as moderate. Studies that reported an overall quality or competency metric (and not specific measures) generally found improvements post-training (Table 3).22, 39, 40, 45, 47, 51, 53, 58, 61, 25, 26, 35, 36, 37 Beyond training completion, 7 non-RCTs reported on CPR quality and AED competency.21, 26, 47, 49, 50, 54, 64 The time points of follow-up varied from 2 months⁴⁷ to 1-year after the initial training.⁶⁴ Recent studies (2012–2020) were able to measure and report on each aspect of CPR skills (Supplementary Table 1. and Table 2).2, 21, 49, 50, 52, 54, 55, 60 Most studies reported an improvement in compression rate/depth or rates being at guidelines standard from baseline skills or immediately post-training (Table 3).49, 50, 52, 54, 55, 60 Correct use of an AED was assessed in one study and showed an improvement from baseline immediately after training.²¹ One study reported retention over time by comparing skills at different time points, they identified that refresher training resulted in less decay of skills (rate, depth, hand position and recoil) over time compared to once-off training.⁵⁴

Table 3.

The impact of BLS training on basic life support skills.

	Outcomes
Basic life support skill	Immediately after training	Six months after training
Chest compression rate	Six studies19, 51, 52, 54, 55, 60 Two reported an improvement from baseline54, 55 Four reported mean rates at guideline standard52, 54, 55, 60 Two reported mean rates under guideline standards19, 51	Three studies49, 50, 54 One reported means rates at guideline standards54 Two reported mean rates under guideline standard49, 50
Chest compression depth	Five studies19, 52, 54, 55, 60 Two reported an improvement in depth from baseline54, 55 Four reported mean depth at guideline standard52, 54, 55, 60 One reported mean depth under guideline standard19	Three studies reported mean depths under guideline standards49, 50, 54
Chest compression fraction	One study reported an improvement from baseline55
Full chest recoil	One study reported a high proportion of compressions with full chest recoil60	Two studies54, 60 One reported a sustained high proportion of compressions with full chest recoil60 One reported no difference54
Hand position during compressions	Four studies reported either a high rate of correct hand position after training52, 60 or an improvement in hand position from baseline54, 55	One study reported a sustained improvement54
Ventilation rate	Three studies reported either an improvement from baseline,54 a ventilation rate at guideline standard21 or a high proportion of successful rescue breaths60	One study reported a sustained improvement from baseline54
Correct AED use	One study reported an improvement in correct AED use from baseline21	One study reported a sustained improvement from baseline, but rate was lower than immediately after training21

AED, automated external defibrillator.

For the outcomes of CPR and AED knowledge at training completion 13 studies (1 RCT,¹⁹ 12 non-RCTs31, 34, 39, 44, 46, 47, 53, 58, 61, 62, 63, 64) were found. Knowledge was often reported using a test created by the study authors. The majority of studies found an increase in knowledge immediately post-training. Only two non-RCTs47, 64 examined knowledge beyond training completion, one at two months⁴⁷ and one at 12 months.⁶⁴ The 12-month study examined the impact of reminders to refresh training, and showed CPR knowledge at 12-months was significantly higher in the two intervention groups (audio-visual and audio-visual-practice training with reminders) compared to control (booklet and DVD with no reminders). No initial post-training assessment was done to assess retention over time.⁶⁴

Confidence

For the important outcome of confidence to perform CPR, we identified five non-RCTs with a low certainty of evidence (downgraded for risk of bias, upgraded for consistency).21, 52, 54, 55, 58 All five studies identified increased confidence following any type of CPR training. However, for studies with ongoing follow-up, a decay in confidence over time was identified.21, 58

Willingness

A moderate certainty of evidence was found from one RCT² for the important outcome of willingness to provide CPR and a low certainty of evidence was found from nine non-RCTs.26, 29, 32, 34, 37, 41, 44, 51, 52 Evidence from the RCT was downgraded to moderate for risk of bias and evidence from non-RCTs was rated as low after being downgraded for risk of bias but upgraded for consistency. Three studies specifically reported willingness before and after training, with all finding a significant increase after training.29, 32, 52 Willingness based on the relationship to the patient was described in two studies, with lower rates as the example theoretical patient (i.e imagine the patient was your father) became less “known” to the participant.26, 32 The method of CPR training was examined in two studies, with one finding slight increases in willingness to perform continuous compression CPR compared to standard CPR², and the other identified traditional “didactic” training to be superior to other forms such as video training.⁴¹

Secondary training

For the new and important outcome of secondary training there was a low certainty of evidence from one RCT² (downgraded for risk of bias) and eight non-RCTs21, 28, 33, 42, 50, 52, 55, 56 (downgraded for risk of bias but upgraded for consistency). These studies describe participants sharing of CPR training and/or teaching materials with others. Of these studies, five reported2, 21, 50, 52, 55 the proportion of participants providing secondary training with rates varying between 22 %⁵⁰ and 72 %.⁵⁵ One study reported 96 % of participants had an intention to teach others, but ultimately only 42 % of participants did with one patient-spouse pair training multiple peers.⁵²

Discussion

Our systematic review examined contemporary literature on the effectiveness of targeting BLS training strategies to family members or carers of people at high-risk of having an OHCA. While we found an additional 17 studies published since 2014 in this updated review, the certainty of evidence remained very low to moderate, and only one recent study reported survival data for instances where BLS skills were used by trained family members.⁵⁹

The included studies that examined the subsequent use of skills and patient outcomes were at high risk of bias. These studies had a high loss to follow-up of participants, which may have included deaths. These studies often used self-reported outcomes and were subject to recall bias. In the larger studies with high follow-up rates, the subsequent number of events and use of skills were low.19, 20, 27, 29 Three studies also reported cases where trained participants were unable to use their skills, as the cardiac arrests occurred either in-hospital or when trained family members were not present.20, 26, 27 Whilst there is a moderate quality of evidence that training improves CPR quality and confidence, it remains unclear if this translates to improved real-world patient outcomes.

Many studies examined skill performance and knowledge immediately following training, with essentially all reporting some improvement post-training. Compared to studies in the previous review, the more recent studies reported objective CPR metrics more frequently, with older studies mostly reporting subjective global skill scores. When global skill scores were reported, they were often created by the authors or adapted from similar studies, with only two studies using a standardised assessment tool.26, 39 The lack of standardised assessment tools or reporting metrics was a major factor precluding any meta-analysis for these outcomes. Future research should focus on reporting objective measurements when reporting skill performance and standardised assessment tools when reporting knowledge to allow inter-study comparisons.

As seen in other trained populations (e.g. health care professionals and general lay public) knowledge, skills and confidence degrade over time highlighting the need to regularly refresh skills.66, 67 Three studies compared outcomes over time following initial training in this review.21, 54, 58 One study found that confidence levels remained unchanged between training and six months post-training.²¹ However, two other studies reported a decline in skills and knowledge over time.54, 58 This highlights the challenge of keeping this cohort competent and confident in their skills. A potential unintended consequence of increased confidence is reduced engagement in retraining, which is essential for preserving BLS skills.⁶⁸ Some studies, such as the feasibility study by Cartledge et al.⁵² send families home with video self-instruction training kits. And while the rate of secondary training was high (a further 87 people were trained), families and high-risk target populations may also benefit from reminders to retrain themselves using the training kits. However, evidence in high-risk populations remains limited. Given that these patients face an ongoing elevated risk of cardiac arrest beyond six months after their initial event,⁶⁹ further research on long-term skill and knowledge retention is crucial.

When training these high-risk populations, there are additional benefits beyond the immediate improvement in skills and knowledge. This includes the potential for high rates of those trained training others in the community (i.e. secondary training). One study reported a secondary training rate as high as 72 % by participants who had an adult family member with underlying cardiovascular disease.⁵⁵ Another and likely unexpected finding was that those empowered with CPR training go on to use their skills to save others in their community. Notably, in the study by Eisenberg et al. the supplied AED was used for seven individuals who were not family members.²⁷ Of these, four achieved return of spontaneous circulation and two survived to hospital discharge. Strategies to enhance secondary training warrant further investigation to expand the potential reach of BLS training.

This review also included a new population, women with a history of illicit drug use who were trained to call for help, administer rescue breathing/CPR and administer naloxone.⁴⁸ In this study, we also see additional training benefits of participants providing secondary naloxone distribution to peers, which has shown to be effective in other studies in reducing overdose mortality.⁷⁰ This demonstrates extended community benefits beyond the intended target population of a high-risk OHCA population.

Limitations and future research

The heterogeneity and poor quality of studies limited this review. In particular, there were few follow-up studies, and sample sizes were often small. Adequately powered studies reporting critical patient outcomes are needed. Long-term follow-up of participants could be strengthened by the use of robust data linkage from hospital or death records, and thus not rely on self-reporting, which is often costly, and prone to loss to follow-up and recall bias. Small sample sizes may be overcome by trialling different participant recruitment timeframes and approaching potential participants later in their recovery trajectory when they can fully consider the benefits of participation.

As the existing evidence demonstrates the benefits of BLS training to high-risk populations, it would likely now be deemed unethical to undertake a future trial that had a control group that received no training (however, long term effects remain unknown). Instead, future research should use accredited CPR/BLS training with standardised and objective reporting of CPR skills and concentrate on the best methods of training (i.e. in person, video self-instruction, hybrid combinations etc.) and retraining. Further, a number of studies reported survival outcomes in a non-standardised manner and thus adherence to (Utstein) reporting guidelines should be encouraged.71, 72

Conclusion

In summary, while our updated systematic review found many recent studies examining BLS training in high-risk groups, the certainty of evidence remained largely low, and few recent studies examined patient outcomes. Despite this, there is continued supporting evidence for improved CPR skills in a simulated setting which increases confidence and willingness to provide CPR, and has the potential for secondary training. Therefore, in light of this evidence, coupled with the fact that a high proportion of OHCAs occur in the home, the evidence continues to support targeting training to family members or carers of those at high risk of OHCA.⁷³

Funding source

JB and SC have been supported by Heart Foundation Fellowships (#104751; #104860) during this work. JF is a NHMRC Investigator Grant recipient (#1174838).

Data availability

Data and template data collection forms can be made available by contacting the authors.

CRediT authorship contribution statement

Zakary Doherty: Writing – original draft, Methodology, Data curation. Janet E. Bray: Writing – review & editing, Supervision, Methodology, Data curation, Conceptualization. Judith Finn: Writing – review & editing, Conceptualization. Susie Cartledge: Writing – review & editing, Supervision, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

The following are the Supplementary data to this article: