Evaluation of culture-specific popular music as a mental metronome for cardiopulmonary resuscitation: a randomised crossover trial

Abstract

Introduction:

Bystander cardiopulmonary resuscitation (CPR) improves survival in out-of-hospital cardiac arrest. The use of certain songs as mental metronomes for CPR have been validated and recognised by contemporary guidelines. We hypothesise that the National Day song, Count on me Singapore (COMS CPR), is not inferior to standard ‘one-and-two-and-three-and-four’ counting (standard CPR) for timing CPR, in terms of the proportion of participants achieving the guideline compression rate of 100–120/minute.

Methods:

This was a prospective randomised crossover trial powered to demonstrate non-inferiority in the CPR rate. After a familiarisation session, volunteers were randomly assigned to two groups. Group A performed one cycle of standard CPR while group B performed one cycle of COMS CPR. Participants then crossed over to perform the other method. The Laerdal SkillReporter measured CPR quality. Four weeks later, participants attended a test scenario, using standard CPR or COMS CPR (randomly allocated).

Results:

Ninety subjects were recruited; 46 were randomly assigned to group A and 44 to group B. Baseline characteristics were similar; 41.1% of COMS CPR achieved 100–120/minute, versus 28.9% of standard CPR (P=0.028). In mixed effects logistical regression, significantly more COMS CPR was performed at 100–120/minute compared to standard CPR (odds ratio 2.44, 95% confidence interval 1.01–5.9, P=0.047). The proportion of insufficient depth was higher in COMS CPR (80.59% vs. 68.01%, P<0.001). There were no differences in other aspects of CPR quality. There were no differences in CPR quality between standard CPR and COMS CPR during the follow-up.

Conclusion:

COMS CPR was not inferior in terms of the proportion of participants delivering a guideline-compliant rate of chest compression. COMS CPR may have applications to layman CPR education, such as in mass education events.

Keywords

Cardiac arrest bystander cardiopulmonary resuscitation basic life support basic cardiac life support

Introduction

Out-of-hospital cardiac arrest (OHCA) is a leading cause of death in Singapore, mirroring the trend in the rest of the developed world. Survival in OHCA hinges on the ‘chain of survival’ concept of Cummins et al., which describes the rapid commencement and seamless provision of a set of rescuer actions.¹ One crucial link in the chain is the early start of cardiopulmonary resuscitation (CPR), which drastically improves survival in OHCA.² As OHCAs frequently occur away from the immediate presence of healthcare providers or proximity to resuscitation facilities, early CPR depends heavily on bystander CPR, which can more than double the patient’s chance of survival.³

In Singapore, the bystander CPR rate was low; 19.7% from 2001 to 2004 and 22.4% from 2010 to 2012.⁴ This was despite having a relatively high proportion of the population (31.4%) trained in CPR (figure based on a national population-based survey in 2010).⁵ The National Resuscitation Council of Singapore has estimated that to achieve a ‘heart-safe’ Singapore, we need to train 500,000 CPR providers in Singapore per annum, ensuring that each household has at least one CPR provider by the year 2020.⁶ Part of the challenge in increasing bystander CPR rates is effective public CPR education. Indeed, studies have found retention of CPR knowledge and skills to decline rather quickly, with some trainees reverting to baseline skills as soon as six months after the training.⁷

Popular songs with tempos that approximate 100–120 beats per minute, which is the compression rate recommended by the 2015 American Heart Association guidelines,⁸ have been variously proposed to be used as metronomes to help time compressions, with some validated as training tools for CPR education. These include the Bee Gees’ Stayin’ Alive⁹ and Ralph Butler’s Nellie the Elephant.¹⁰ Studies have found varying suitability for application to CPR training. One study compared timing CPR using a popular Korean song with using a metronome for real-time auditory feedback and found that the subjects randomly assigned to be taught using the popular song achieved a better guideline-compliant compression rate six months after the training.¹¹ Hafner et al. similarly found rate compliance at six weeks to be better with Stayin’ Alive.¹² While useful in certain communities, the effectiveness of these songs as a CPR training adjunct is limited by the familiarity of the target population with the song.

Pilot experience with using the chorus of a national education song in Singapore, Count on me Singapore (COMS) by Clement Chow (see Appendix 1 for lyrics), as a mental metronome for training laypersons was well received and has found favourable results in resultant CPR quality. We hypothesised that COMS CPR is non-inferior to standard CPR in achieving a guideline-recommended compression rate of 100–120/minute, with no adverse impact on CPR quality, measured by the depth of compression, correct hand placement and adequate recoil. We also hypothesised that COMS CPR causes a lower fatigue level and higher satisfaction in providers after performing one cycle of CPR, as singing/humming the song provided breaks as opposed to continuous counting.

Methods

Setting

Singapore is an urbanised, island city-state in South-east Asia with a population of 5.6 million and a high population density over 715 square kilometers.¹³ Emergency medical services in Singapore are publicly funded and provided by the Singapore Civil Defence Force.¹⁴ A national population-based survey found 31% of the population to have had prior training in CPR, and 11% in automated external defibrillation (AED) application.⁵ Utstein survival for OHCA between 2010 and 2012 was 11.0%.⁴ Median door-to-balloon time for ST-segment elevation myocardial infarction was 64 minutes between 2010 and 2012.¹⁵ Singapore has adopted a dispatcher-assisted CPR protocol since 2012.¹⁶

The Singapore Armed Forces is the military component of the Ministry of Defence in Singapore. The Singapore Armed Forces medical corps aims to ensure excellent, comprehensive healthcare for our soldiers and optimise their combat performance.¹⁷ Since 2010, all enlisted recruits receive training in CPR and AED application during the first phase of their military training. Recruits who are subsequently assigned medical vocations receive additional systematic training accredited by the Justice Institute of British Columbia, to provide first-line treatment for a range of medical emergencies including OHCA.

Study design

This was a prospective, randomised, crossover non-inferiority trial comparing CPR timed using Count on me Singapore (COMS CPR) and CPR timed using the standard ‘one-and-two-and-three-and-four’ (standard CPR). A simple coin-based randomisation procedure randomly assigned subjects to either group A or group B. Group A would first perform one cycle (two minutes) of standard CPR, while group B would first perform one cycle of COMS CPR. After a 15-minute rest interval, the subjects would cross over to perform one cycle of the other method of CPR. After completing this second cycle, a survey form was administered to the subjects. After a four-week period, subjects were recalled to attend a test scenario, by way of coin-based simple randomisation.

All CPR was performed on the Laerdal Resusci-Anne SkillReporter manikin (Laerdal Medical Corporation, Stavanger, Norway). Participants were not allowed to view the SkillReporter meter during CPR for real-time visual feedback. Due to practical differences in the two methods of CPR, there was no blinding of the participants and observers to the allocated treatment. However, the data analyst was blinded to treatment allocation. The recruitment and trial protocol were administered over three days in October 2016 in a military camp medical centre, with a follow-up test after a two-week interval.

Study population

Members of an infantry battalion were invited to enroll in the study. This included members from the enlisted, specialist and officer ranks. They were recruited by invitation issued through the battalion management. The exclusion criteria from the trial were: (a) existing medical conditions that may cause danger to the participant (e.g. cardiac or musculoskeletal conditions); and (b) having attended a recent CPR course (within one month). The latter group was excluded as it would confer advantage to the standard CPR group.

Interventions

The song investigated, Count on me Singapore by Clement Chow, is taught as part of national education in all non-international primary schools in Singapore as far back as the 1980s. Regular playbacks on mass media during the yearly National Day period also provide the public with regular refreshers. The tempo of the chorus as sung by participants of the pilot study ranged from 105 to 124 per minute (median 112/minute). Each chorus of the song approximated 32 beats. The lyrics are sung mostly on the ‘on’ beats while ‘off’ beats are silent, allowing performers to catch their breath, hence COMS CPR is hypothesised to be less tiring. After informed consent was obtained (during which volunteers were informed on the procedure of COMS CPR), participants underwent a standardised 15 minutes instructional session involving an instructional video explaining COMS CPR.

Variables and outcome measures

The primary outcome was the proportion of participants achieving a guideline-compliant rate (100–120/minute) of chest compression. A null hypothesis was assumed – that COMS CPR was inferior to standard CPR in the primary outcome.

Secondary outcomes were: (1) Other CPR quality measures, namely (a) proportion of participants achieving adequate mean depth of compression (5 cm), (b) proportion of compressions with incorrect hand placement, (c) proportion of compressions with inadequate recoil; (2) Fatigue level after a cycle of CPR (two minutes) on a 10-point visual scale; (3) Satisfaction of CPR provider on a five-point Likert scale in terms of (a) ease of learning, (b) fatigue, (c) overall effectiveness; and (4) Ability to retain compression rate after a four-week interval with no additional refresher in between. Quantitative CPR quality variables were collected from reports printed from the Laerdal Resusci-Anne SkillReporter. Data on demographics and previous CPR training were collected using an anonymised data collect sheet.

Sample size calculation

Based on previous studies that reported a standard deviation of 13.125 for a paired difference in mean compression rate, it was estimated that 64 subjects would be required to detect a mean paired difference between any two interventions of five compressions per minute with a power of 0.85 and a type I error, alpha of 0.05.¹⁰ We aimed to recruit 80 (rounded up to account for drop-outs) volunteers, i.e. 40 volunteers in sequence (COMS CPR, standard CPR) and 40 volunteers in sequence (standard CPR, COMS CPR).

Statistical analysis

Frequency tables and descriptive statistics for all outcome variables listed above were calculated. Baseline characteristics were compared between the standard CPR and COMS CPR groups. The chi-square or McNemar tests were used to compare categorical or binary factors of CPR quality measures between standard CPR and COMS CPR. The paired t-test was used to compare normally distributed continuous outcomes of CPR quality measures and the signed rank test was used to assess the continuous outcomes which are not normally distributed. Carry-over, sequence and period effects were measured using an analysis of variance (ANOVA) model in the specific STATA procedure (command pkcross) for cross-over design experiments. Sequence is defined as either sequence 1 (standard CPR ⩾ COMS CPR) or sequence 2 (COMS CPR ⩾ standard CPR), period as period 1 (after random assignment and before crossover) or period 2 (after crossover), Figure 1. The performance of standard CPR and COMS CPR after two weeks was assessed using the independent t-test or the Mann–Witney U test when appropriate.

Figure 1.

Population flow diagram.

The independent relationship between standard CPR and COMS CPR and CPR quality outcomes after adjusting for individual characteristics and ensuring there was no carry over, sequence or period effects was tested. These were tested using mixed-effects linear, logistic and ordered logistic regression models to account for the correlation of repeated response over time in this crossover trial, as appropriate depending on the data type of the CPR quality outcome. We selected the included variables and outcomes in the multivariate analyses based on whether they were significant at P<0.1 in the univariate analysis and used a backward step-wise procedure in the multivariable models. Data were analysed using STATA version 14.¹⁸ Statistical significance was set at P<0.05.

Ethics approval/Clinical Trials registration

The central institutional review board (CIRB) at the Singapore General Hospital granted approval for this research project with a waiver of patient consent (CIRB reference: 2016/2773). Operational approval and additional ethics review were obtained from the study site, Singapore Armed Forces. This study was registered with ClinicalTrials.gov, no. NCT02940964.

Results

Population flow and baseline characteristics

The trial recruited 90 subjects. None of the subjects met any exclusion criteria; 46 subjects were randomly assigned to group A and 44 to group B. A CONSORT flow diagram is presented in Figure 1. All 90 subjects completed the first phase of the trial and were analysed in their respective allocated groups. Forty-four (48.9%) subjects did not attend the test held two weeks after the first phase. Table 1 shows the baseline characteristics of the subjects. Subjects in both groups A and B were similar in terms of demographics and prior CPR training. The median age was 21.3 years, and 100% were men.

Table 1.

Baseline characteristics of study participants in the COMS CPR and standard CPR arms.

Characteristic	Group A (sequence 1) (n=46)	Group B (sequence 2) (n=44)
Age (years), median (IQR)	21.2 (19.9–21.8)	21.3 (20.4–22.1)
Men, n (%)	46 (100)	44 (100)
Race, n (%)
Chinese	35 (76.1)	32 (72.7)
Malay	7 (15.2)	5 (11.4)
Indian	3 (6.5)	3 (6.8)
Others	1 (2.2)	4 (9.1)
Body mass index (kg/m²) median (IQR)	21.9 (19.8–22.9)	21.9 (20.1–24.2)
Highest education level, n (%)
‘N’ levels or below	2 (4.3)	2 (4.5)
‘O’ levels	2 (4.3)	2 (4.5)
‘A’ levels or diploma	41 (89.1)	35 (79.5)
Bachelor’s degree	0 (0)	5 (11.4)
Higher than bachelor’s	1 (2.2)	0 (0)
Healthcare training, n (%)
Non-healthcare provider	35 (79.5)	36 (83.7)
EMT trained	9 (20.5)	6 (14)
Nursing trained	0 (0)	1 (2.3)
Prior CPR training, n (%)
No prior training	1 (2.2)	2 (4.8)
Expired CPR training	17 (37.8)	14 (33.3)
Current CPR training	27 (60)	26 (61.9)
Prior AED training, n (%)
No prior training	2 (4.4)	4 (9.3)
Expired AED training	18 (40)	12 (27.9)
Current AED training	25 (55.6)	27 (62.8)
Has CPR experience on real patient, n (%)	1 (2)	0 (0)

AED: automated external defibrillator; COMS: Count on me Singapore; CPR: cardiopulmonary resuscitation; EMT: emergency medical technician (medic); IQR: interquartile range.

Primary outcomes

In the context of our primary outcome (the proportion of subjects who achieved a chest compression rate of 100–120/minute), 41.1% of COMS CPR performed achieved 100–120/minute, versus 28.9% of standard CPR (P=0.028, Table 2). In mixed effects logistical regression, significantly more COMS CPR was performed at 100–120/minute compared to standard CPR (odds ratio (OR) 2.44, 95% confidence interval (CI) 1.01–5.9, P=0.047). This model was adjusted for age, gender, ethnicity, education level, healthcare training, prior CPR training, prior AED training, previous CPR experience on real patients, body mass index (BMI) and period effect.

Table 2.

Comparison of quality of CPR measures between COMS CPR and standard CPR arms.

Total 90 participants	Standard CPR	COMS CPR	P value	Effect,^a P value
				Sequence	Carry-over	Period
Compression rate	130.06 (14.72)	122.9 (14.76)	<0.001*	0.253	0.146	0.001*
Compression depth	42 (9.91)	38.69 (10.1)	0.002*	0.173	0.087	<0.001*
% of insufficient depth	68.01 (37.3)	80.59 (32.8)	<0.001*	0.597	0.508	<0.001*
% of insufficient recoil	2.8 (12)	5.5 (15.9)	0.101	0.897	0.879	0.417
% of incorrect hand placement	11.39 (24.2)	17.55 (32.54)	0.146	0.349	0.341	0.143
% of participants achieving 100–120 (bpm), n (%)	26 (28.9)	37 (41.1)	0.028*	0.401	0.318	0.028*
% of participants achieving >100 (bpm), n (%)	89 (98.9)	86 (95.6)	0.083	0.735	0.688	0.082
Multivariable analysis		OR (95% CI)	P value
% of participants achieving 100–120 (bpm)^b	Reference	2.44 (1.01–5.9)*	0.047*

Data presented in the table are mean (standard deviation) unless specified.

P value for sequence, carry-over and period effect using analysis of variance for a crossover study.

Adjusted for age, gender, ethnicity, education level, healthcare training, prior CPR training, prior AED training, any CPR experience on real patients, BMI.

P<0.05 (statistically significant).

AED: automated external defibrillator; COMS: Count on me Singapore; CPR: cardiopulmonary resuscitation; BMI: body mass index; bpm: beats per minute; OR: odds ratio; CI: confidence interval.

Secondary outcomes

The proportion of insufficient depth was higher in the COMS CPR group (80.6% vs. 68.0%, P<0.001). There were no significant differences in other aspects of CPR quality studied (proportion of insufficient recoil and proportion of incorrect hand placement). There were no significant differences in the median fatigue score immediately after two minutes of CPR, and survey scores on overall effectiveness, ease of learning and how tiring (Table 3). A breakdown of these survey scores can be found in Table 3. During the unannounced follow-up test four weeks later, there were no significant differences between standard CPR and COMS CPR in compression rate, depth, proportion of insufficient recoil, proportion of incorrect hand placement or tiredness scale (Table 4).

Table 3.

Comparison of satisfaction and fatigue scores between COMS CPR and standard CPR.

Total 90 participants	Standard CPR	COMS CPR	P value
Fatigue score immediately after 2 minutes of CPR (0–10), mean (standard deviation)	4.77 (1.64)	4.51 (1.8)	0.129
Survey
Overall effectiveness (1–5) n (%)			0.845
2	4 (4.4)	2 (2.2)
3	22 (24.4)	24 (27)
4	45 (50)	43 (48.3)
5	19 (21.1)	20 (22.5)
Ease of learning (1–5) n (%)			0.117
2	0 (0)	3 (3.4)
3	16 (17.8)	17 (19.1)
4	38 (42.2)	26 (29.2)
5	36 (40)	43 (48.3)
How tiring (1–5) n (%)			0.465
1	4 (4.4)	6 (6.7)
2	17 (18.9)	16 (18)
3	44 (48.9)	33 (37.1)
4	23 (25.6)	30 (33.7)
5	2 (2.2)	4 (4.5)
What do you think should be taught to new recruits? n (%)	11 (12.22)	10 (11.11)	NA

COMS: Count on me Singapore; CPR: cardiopulmonary resuscitation.

Table 4.

Comparison of standard CPR versus COMS CPR performance during an unannounced follow-up test 7–14 days later.

Total 46 participants	Standard CPR	COMS CPR	P value
Compression rates	120.79 (17.07)	125.05 (12.04)	0.338
Compression depth	48.5 (5.89)	46.18 (7.58)	0.251
% of insufficient depth	53.98 (38.42)	51.86 (39.25)	0.854
% of insufficient recoil	5.65 (14.03)	3.86 (8.39)	0.607
% of incorrect hand placement	33.73 (41.8)	21.74 (27.98)	0.264
Tiredness scale (0–10)	5.5 (1.77)	5.27 (1.67)	0.657

Data presented in the table are mean (SD) unless specified.

COMS: Count on me Singapore; CPR: cardiopulmonary resuscitation.

Discussion

In the population studied, CPR timed using a popular song Count on me Singapore was not inferior to that timed using the conventional ‘one-and-two-and-three-and-four’ method, in terms of the proportion of subjects achieving a guideline-compliant rate of chest compression (100–120/minute).⁸ However, it was associated with a higher proportion of subjects not achieving sufficient depth of compression.

These findings echo previous studies that tested popular songs as mental metronomes for CPR. Rawlins et al. examined the use of the song That’s the Way (I Like It) on 130 participants and found that subjects randomly assigned to be taught using the song were more likely to provide CPR at an appropriate rate, but also that as the compression rate increases, the proportion of compressions delivered at the correct depth reduces.¹⁰

Our result of a lower proportion of subjects achieving adequate compression depth is an important one. Based on a similar finding, Rawlins et al. had suggested that the songs they studied ought not to be used for CPR.¹⁰ The issue may warrant a deeper analysis. The shallower depth may not entirely be caused by the use of the songs. In fact, the shallower depth of compression with an increasing rate is not a finding unique to these studies. In the absence of guidance from high quality trials, our understanding of the ideal compression rate stems from large observational studies and animal studies.¹⁹ In two important studies analysing OHCA cases from the Resuscitation Outcomes Consortium, Idris et al. found that the probabilities of return of spontaneous circulation and survival were highest when the compression rate was 100–120/minute, even after adjusting for depth. The same study also found depth to decrease greatly as the rate increased. This gives credence to the observation that despite being associated with shallower depth, a guideline-compliant rate still improved outcomes. A possible explanation for the reduced depth is that the intervention group was focused on achieving the correct rate, as briefed by the trial administrators, at the expense of achieving the correct depth. We propose that adequate depth needs to be emphasised to trainees if popular songs are to be used.

The randomised crossover design allowed each subject to act as his own control and reduced the effect of the obvious confounder that the vast majority of the study population had previous training in standard CPR. All subjects completed the first phase to contribute data for the primary outcome variable.

The study design lends itself to several limitations. First, the study population of mainly healthy young men may not be representative of the general population in terms of fitness. As an example, the BMI of the performer has previously been shown to affect CPR quality.²⁰ Second, the recruitment process which invited volunteers may select for subjects with a keener interest in CPR or in the song. Third, due to fundamental differences in the two methods of CPR, neither the subject nor the trial administrator could be blinded. Fourth, the simulated setting of this trial may only approximate real-life circumstances when factors such as psychological stress or physical barriers may be involved.¹⁶ Finally, the turnout rate for the second phase of the trial (consisting of a test two weeks after phase 1) was low at approximately 50%. This may overestimate the CPR quality of COMS CPR during the test as the subjects who turned up may have been more motivated.

CPR improves survival in OHCA and can be effectively taught to most laypersons.²¹ However, it is challenging to prevent the decline of acquired CPR skills in trained laypersons. In considering strategies for mass CPR education, considerations beyond CPR quality alone are important. An advantage of a CPR education method that hinges on the familiarity with a (national education) song among the general population is that they may be motivated to learn and perform CPR. The poor retention rate of CPR training found by previous studies⁷ may be partly attributed to a lack of real-life practice after the training, and hence disuse. There is possible benefit in relating CPR training to concepts that the general population can periodically revise. The national education theme also allows for opportunistic teaching (including refresher) events such as mass CPR education events on National Day. The use of a musical metronome that is ubiquitous and easily utilised presents the potential for ambulance dispatch centres to guide callers through CPR if the caller has previously been taught COMS CPR and can relate more readily to it than standard CPR. Finally, the chorus length of 32 beats allows for the delivery of ventilation between each chorus, hence simplifying CPR and reducing the learning curve.

COMS CPR would to some extent address issues relating to retention of the skill currency of bystanders. However, there are also other factors contributing to a low bystander CPR rate, including those related to recognition of the arrest and reluctance due to concerns about hygiene associated with mouth-to-mouth ventilation.⁵ Therefore, to increase the critical mass of the first responder network for OHCA, adjuncts to improve CPR education of laypersons must continue to be studied for their effectiveness.

Conclusion

Compared to standard CPR, COMS CPR was not inferior in terms of the proportion of participants delivering a guideline-compliant rate of chest compression. However, attention should be given to achieving sufficient depth. COMS CPR may have applications to layman CPR education, such as in mass education events and schools.

Footnotes

Appendix

Appendix 1.

Lyrics to chorus of Count on me Singapore (dots denote on beats for chest compression).

1 and	2 and	3 and	4 and	5 and	6 and	7 and	8 and
Count	on	me	Singapore
1 and	2 and	3 and	4 and	5 and	6 and	7 and	8 and
Count	on	me	Singapore
1 and	2 and	3 and	4 and	5 and	6 and	7 and	8 and
Count	on	me	to give my best		and	more
1 and	2 and	3 and	4 and	5 and	6 and	7 and	8 and
Count	on	me	Singapore

Lyrics reproduced with permission of the Government of Singapore as represented by the Ministry of Culture, Community and Youth.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Availability of data and materials

The datasets generated and/or analysed during the current study are available from AFWH.

Authors’ contributions

AFWH and ZL researched the literature and conceived the study. AFWH, ZL, RMT and MEHO were involved in protocol development, gaining ethical approval, patient recruitment and data analysis. AFWH wrote the first draft of the manuscript. All authors reviewed and edited the manuscript and approved the final version of the manuscript.

Conflict of interest

The authors declare that there is no conflict of interest.

Informed consent

Written informed consent was obtained from the patients for their anonymised information to be published in this article.

Ethical approval

The central institutional review board at the Singapore General Hospital granted approval for this research project with a waiver of patient consent (CIRB reference: 2016/2773). Operational approval and additional ethics review were obtained from the study site, Singapore Armed Forces.

Trial registration

, no. NCT02940964.

References

Cummins

Ornato

Thies

, et al. Improving survival from sudden cardiac arrest: the “chain of survival” concept. Circulation 1991; 83: 1832–1847.

Cummins

Eisenberg

Hallstrom

, et al. Survival of out-of-hospital cardiac arrest with early initiation of cardiopulmonary resuscitation. Am J Emerg Med 1985; 3: 114–119.

Rea

Eisenberg

Culley

, et al. Dispatcher-assisted cardiopulmonary resuscitation and survival in cardiac arrest. Circulation 2001; 104: 2513–2516.

Lai

Choong

Fook-Chong

, et al. Interventional strategies associated with improvements in survival for out-of-hospital cardiac arrests in Singapore over 10 years. Resuscitation 2015; 89: 155–161.

Ong

MEH

Quah

JLJ

AFW

, et al. National population based survey on the prevalence of first aid, cardiopulmonary resuscitation and automated external defibrillator skills in Singapore. Resuscitation 2013; 84: 1633–1636.

Anantharaman

Developing resuscitation programmes in the community: the tasks ahead for the National Resuscitation Council. Singapore Med J 2011; 52: 634–641.

Weaver

Ramirez

Dorfman

, et al. Trainees’ retention of cardiopulmonary resuscitation. JAMA 1979; 241: 901.

Neumar

Shuster

Callaway

, et al. Part 1: Executive summary: 2015 American Heart Association guidelines update for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation 2015; 132 (Suppl. 2): S315–S367.

Hafner

Sturgell

Matlock

, et al. Stayin’ alive: a novel mental metronome to maintain compression rates in simulated cardiac arrests. J Emerg Med 2012; 43: e373–e377.

10.

Rawlins

Woollard

Williams

, et al. Effect of listening to Nellie the Elephant during CPR training on performance of chest compressions by lay people: randomised crossover trial. BMJ 2009; 339: b4707.

11.

Hong

Hwang

Lee

, et al. Metronome vs. popular song: a comparison of long-term retention of chest compression skills after layperson training for cardiopulmonary. Hong Kong J Emerg Med 2016; 23: 145–152.

12.

Hafner

Jou

Wang

, et al. Death before disco: the effectiveness of a musical metronome in layperson cardiopulmonary resuscitation training. J Emerg Med 2015; 48: 43–52.

13.

Department of Statistics, Singapore. Population and population structure. www.singstat.gov.sg/find-data/search-by-theme/population/population-and-population-structure/latest-data (2018, accessed 9 December 2018).

14.

AFW

Chew

Wong

, et al. Prehospital trauma care in Singapore. Prehosp Emerg Care 2015; 19: 409–415.

15.

AFW

Loy

Pek

, et al. Emergency medical services utilization among patients with ST-segment elevation myocardial infarction: observations from the Singapore Myocardial Infarction Registry. Prehospital Emerg Care 2016; 3127: 1–8.

16.

AFW

Sim

Shahidah

, et al. Barriers to dispatcher-assisted cardiopulmonary resuscitation in Singapore. Resuscitation 2016; 105: 149–155.

17.

Ministry of Defence, Singapore. Singapore Armed Forces Medical Corps. www.mindef.gov.sg/imindef/mindef_websites/atozlistings/safmc/home.html (2015, accessed 10 January 2017).

18.

StataCorp. Stata Statistical Software: Release 14. College Station, TX: Statacorp, 2015.

19.

AFW

Pek

Yap

, et al. Compression rate during cardiopulmonary resuscitation. J Korean Med Sci 2016; 31: 1851–1852.

20.

Hong

Park

Lee

, et al. A different rescuer changing strategy between 30:2 cardiopulmonary resuscitation and hands-only cardiopulmonary resuscitation that considers rescuer factors: a randomised cross-over simulation study with a time-dependent analysis. Resuscitation 2012; 83: 353–359.

21.

Cummins

Eisenberg

MS.

Prehospital cardiopulmonary resuscitation. Is it effective?

JAMA 1985; 253: 2408–2412.