Secondary Logo

Journal Logo

Original Articles

Change in Cardiopulmonary Resuscitation Performance Over Time During Simulated Pediatric Cardiac Arrest and the Effect of Just-in-Time Training and Feedback

Duff, Jonathan P. MD; Bhanji, Farhan MD; Lin, Yiqun PhD; Overly, Frank MD§; Brown, Linda L. MD, MSCE§; Bragg, E. Alexis MD; Kessler, David MD, MSc; Tofil, Nancy M. MD#; Bank, Ilana MD; Hunt, Elizabeth A. MD, PhD††; Nadkarni, Vinay MD, MS∗∗; Cheng, Adam MD; for the INSPIRE CPR Investigators

Author Information
doi: 10.1097/PEC.0000000000002359
  • Free


Effective chest compressions (CCs) are key to survival of any patient having cardiac arrest. High-quality cardiopulmonary resuscitation (CPR) has been directly related to survival and improved neurological outcome after cardiac arrest in both children and adults.1–6 The 2015 American Heart Association CPR guidelines identified components of high-quality CPR including maintaining adequate compression depth and rate and minimizing excessive leaning force and no-flow time.7

Despite this increased focus on maintaining CC quality, numerous studies in both simulated environments and real patients have shown the quality of CC provided remains poor.8–10 As CPR guidelines have changed to stress minimizing breaks in CPR and maintaining adequate depth (5–6 cm) and rate (100–120 CC/min), rescuer fatigue may play a role in the poor quality of CC provided by rescuers. Simulation-based studies have shown a drop off in quality of CC, especially with longer resuscitation times.11–21

Most of the studies to date have focused on CC quality over time delivered by a single rescuer performing CC-only CPR on an adult task trainer. There have been pediatric studies, but most of these have also focused solely on single rescuers, both with and without rescue breathing.14,21,22 It is unclear if a similar drop off in CPR quality would occur in a team-based resuscitation setting, with multiple rescuers providing CC under the direction of a team leader.

In this study, we aim to describe the effect of prolonged resuscitation on CPR quality performed by 2 providers, as measured by CC depth and rate, within a 12-minute simulated cardiac arrest. We have previously examined the effects of 2 interventions on CC quality, visual feedback, and just-in-time (JIT) training.23 In this study, we examined the effects of these interventions on any decay in CC quality over a prolonged pediatric simulated resuscitation.


This study is a secondary analysis of data collected from the CPRCARES study.23 The CPRCARES study was a multicenter, prospective, randomized, controlled trial to assess the quality of CCs performed during simulated pediatric cardiac arrest. Research ethics board approval was obtained from each participating institution, and informed consent was obtained from all participants. Participants were assigned to 3-person teams: a designated team leader and 2 team members serving as CPR providers. Team leaders were senior-level residents in pediatrics, anesthesia, emergency medicine, or fellows in pediatric critical care, pediatric emergency medicine, or pediatric anesthesia. Team members could be nurses, medical students, or residents. During the simulation scenarios, they were joined by 2 scripted confederates to form a 5-person team. One confederate worked as a medication nurse and aided with calculating drug doses. Only the 2 study participants acting as team members provided CCs during the 12-minute scenario; the team leader and confederates did not perform CC at any time. The timing of when to switch compressors was up to each individual resuscitation team.

The study examined the effects of 2 interventions: (a) JIT CPR training with visual feedback during CPR practice and (b) use of visual feedback device during the simulated cardiac arrest. The Laerdal CPRCard was used to provide corrective feedback on CC rate and depth. Details of the initial study design have been reported elsewhere.23 In this study, we report results from participants recruited to the control arm (no intervention), visual feedback–only arm, and JIT training–only arm. This was done for ease of analysis and to focus on any decay in performance over time, rather than on the original interventions.

In the visual feedback arm (the only arm in which feedback from the device was visible), study participants were oriented to the feedback from the device and how to interpret the data. In the JIT arm, participants viewed a short training video on CPR technique and the importance of switching CC providers every 2 minutes.

After an orientation to the study and the simulation environment, each team participated in 2 simulated pediatric resuscitations. The first was a septic shock scenario with no CPR required to prevent the participants from artificially fixating on CPR during the study. The second scenario was a pediatric cardiac arrest with an intubated 5-year-old patient, starting in pulseless electrical activity and progressing to ventricular fibrillation. The simulation environment and each scenario were standardized across sites to minimize threats to internal validity.24 The Laerdal SimJunior mannequin was used for all scenarios.

Cardiopulmonary resuscitation quality data were collected from the Laerdal SimJunior mannequin including CC depth and rate. Data were collected in 30-second epochs, which were then grouped into 3 different time intervals: first 4 minutes, middle 4 minutes, and final 4 minutes of the simulated cardiopulmonary arrest event. We did not track when the resuscitation teams swapped compressors. They were encouraged to swap compressors every 2 minutes as per the American Heart Association guidelines.7 Therefore, a 4-minute interval was chosen a priori to represent an interval most likely to include both team members providing CCs, not just one.

Outcome Measures

The primary outcome measure was mean compression depth in each of the 3 different 4-minute intervals. Depth was chosen as the primary outcome as it is more closely correlated with patient outcomes.4,25 The secondary outcome was mean compression rate in each of the 3 different 4-minute intervals.

Sample Size

We used study subjects from the CPRCARES study as a convenience sample for this study. Details of sample size calculation are reported elsewhere.23

Statistical Analysis

All data analysis was performed using R version 3.2.0. We used 30-second epochs of CC performance as the unit of measure for depth and rate of CPR. We used descriptive statistics (mean and SD) to describe the mean compression depth and mean compression rate within 0 to 4 minutes (first interval), 4 to 8 minutes (middle interval), and 8 to 12 minutes (final interval) of cardiopulmonary arrest scenarios in each study arm. Repeated-measure analyses of variance were conducted to determine if the mean compression depth and rate differed in the 3 intervals during the cardiopulmonary arrest scenario.


Study Population

Sixty-seven events, including 134 CPR providers, were analyzed for the study (Table 1). One site from the main study was excluded due to incomplete data. An additional 5 events from the no intervention group, 4 events from the visual feedback group, and 5 events from the JIT training group were excluded due to technical issues resulting in incomplete data. A total of 1605 epochs from 67 events were included in the analysis, with 528 epochs (22 events) in the no intervention group, 552 epochs (23 events) in the visual feedback group, and 525 epochs (22 events) in the JIT training group. The demographic characteristics of the participants across sites are reported in Table 1.

TABLE 1 - Participant Demographics
Characteristics Group No Intervention Visual Feedback JIT Training
Sex, n (%) Female 34 (77.3) 32 (69.6) 31 (70.5)
Male 10 (22.7) 14 (30.4) 13 (29.5)
Occupation, n (%) Nurse 21 (47.7) 24 (52.2) 25 (56.8)
Attending pediatrician 1 (2.3) 1 (2.2) 0 (0)
Medical student 6 (13.6) 8 (17.4) 5 (11.4)
Resident 16 (36.4) 13 (28.3) 14 (31.8)
Last BLS/ACLS/PALS course taken, n (%) <6 mo 16 (36.4) 25 (54.3) 21 (47.7)
6–12 mo 18 (40.9) 12 (26.1) 16 (36.4)
>12 mo 10 (22.7) 9 (19.6) 7 (15.9)
CPR done on simulator or real patients in the past 2 y, n (%) 1–5 times 32 (72.7) 33 (71.7) 31 (70.5)
>5 times 3 (6.8) 4 (8.7) 4 (9.1)
Never 9 (20.5) 9 (19.6) 9 (20.5)
Total 44 46 44
BLS indicates basic life support; ACLS, advanced cardiac life support; PALS, pediatric advanced life support.

No Intervention

The mean compression depths and rates in the 0- to 4-, 4- to 8-, and 8- to 12-minute intervals are shown in Table 2 and Figures 1, 2. The mean compression rates in the no intervention group for the first, middle, and final intervals were 128.2/min, 129.9/min, and 132.8/min, respectively. There was a statistically significant increase in rate across the 3 intervals (P = 0.043). The mean compression depths in the no intervention group for the 3 intervals were 32.0, 31.8, and 31.5 mm, respectively. The difference between 3 intervals was not statistically significant (P = 076).

TABLE 2 - Effect of Visual Feedback and JIT Training on Quality of Chest Compressions Over Time
0–4 Min 4–8 Min 8–12 Min P
CC depth, mean (SD) No intervention 32.0 (8.6) 31.8 (8.2) 31.5 (7.9) 0.761
Visual feedback 36.2 (7.2) 36.7 (7.4) 36.2 (7.2) 0.704
JIT training 39.7 (8.1) 39.0 (9.2) 38.1 (8.1) 0.066
CC rate, mean (SD) No intervention 128.2 (18.8) 129.9 (21.5) 132.8 (19.7) 0.044
Visual feedback 115.1 (9.8) 116.8 (12.2) 116.4 (9.4) 0.447
JIT training 115.9 (12.1) 114.5 (18.7) 117.6 (15.7) 0.111

The effect of no intervention (left), visual feedback (middle), and JIT training (right) on CC depth (in millimeters) across each time interval.
The effect of no intervention (left), visual feedback (middle), and JIT training (right) on CC rate across each time interval.

Visual Feedback

In the group that received visual feedback only, the mean compression rates in the first, middle, and final intervals were 115.1/min, 116.9/min, and 116.4/min, respectively. The mean compression depths in the 3 intervals were 36.2, 36.7, and 36.3 mm, respectively. There was no significant change in depth (P = 0.704) or rate (P = 0.447) across the 3 time intervals.

Just-in-Time Training

In the group that received JIT CPR training only, the mean compression rates in first, middle, and final intervals were 115.9/min, 114.5/min, and 117.6/min. The mean compression depths in 3 intervals were 39.7, 39.0, and 38.1 mm, respectively. There were no significant changes in depth (P = 0.066) or rate (P = 0.111).


We demonstrated that there is no significant change in quality of CPR provided by 2 rescuers during a 12-minute simulated pediatric cardiac arrest. There was no significant change in depth across the 12 minutes in any of the 3 study arms. There was a significant increase in rate in the group with no JIT or visual feedback (from 128.2/min to 132.8/min). It is unclear as to why CC rate increased over the study, although the rapid rate is concerning. Quality of CC provided was often not guideline compliant in all 3 arms of the study, consistent with published literature on real pediatric arrests.9,26 This study is the first to describe CC quality over time in a realistic pediatric cardiopulmonary arrest scenario with 2 rescuers, rather than a single rescuer performing CC on a simple task trainer. The complexity of the resuscitation is important as it has been shown that in simulated scenarios involving multiple tasks, CC quality is affected.27

Other pediatric studies examining CC performed by single rescuers on task trainers have shown deterioration in CC quality over time. Li et al22 asked Neonatal Resuscitation Program (NRP) providers to perform compressions at a 3:1 compression/ventilation ratio on a neonatal mannequin (with a confederate providing ventilations). They noted a significant reduction in depth of CC provided within 3 minutes of the start of the scenario. In a pediatric task trainer study, it was shown that single rescuers were only able to maintain guideline-compliant CC depth (≥38 mm at the time of the study) for 3 minutes of continuous compressions.21 They noted an increase in rate over the simulated resuscitation as we did in our study.

In our study, the resuscitation team was allowed to swap compressors at their own discretion, which may explain why we did not see a decline in depth over the course of the resuscitation. However, it is also possible that the reason we did not see a drop in quality over time was because the initial quality of CC provided was so poor. In a study of paramedics providing CC, the authors noted that rescuers that provided poor-quality CC at the beginning of the study did not demonstrate significant variability over the course of the simulated resuscitation.28 Only a proportion of those participants that started with CC that were meeting guidelines showed evidence of decay over time.

A number of techniques have been put forward as methods to improve quality of CC provided. Devices that provide immediate feedback to rescuers providing CC have been shown to improve CC quality in simulated23 and real9 pediatric cardiac arrests. More work is needed to determine their effect on CC quality over time. Chung et al29 found that an audio metronome only worsened depth in simulated cardiac arrest and did not prevent a drop in depth with time. In a study of adult patients with cardiac arrest, Sugerman et al30 noted a significant decrease in depth over time, despite the use of an audio feedback device. In our study, neither the visual feedback back arm nor the control arm demonstrated a significant decline in depth over the 12-minute arrest.

Just-in-time training is another technique that has been shown to improve the acquisition and retention of CPR skills.31,32 No prior research has explored the effect of JIT training on provider fatigue during management of cardiopulmonary arrest. We speculate that the lack of deterioration in CPR quality over time in all 3 groups may be related to the fact that participants in all groups had an opportunity to practice CPR for 2 minutes just before the simulated event. The JIT training group differed in that they watched a short CPR training video and were provided real-time CPR visual feedback during their JIT practice session. Future work should clearly delineate the effect of JIT training on rescuer fatigue by comparison with a group without any prior CPR practice.

Study Limitations

This study has a number of limitations. The simulation scenario was only 12 minutes long, which, with 2 rescuers available to provide compressions, may not be long enough to demonstrate deterioration in CPR quality. It is always difficult to know if data from the simulated environment can translate to real patients—providers may perform differently with a real patient. The team composition was similar in the 3 arms but may not be generalizable to all CPR teams. There are a number of variables that may affect the quality of CC a rescuer can provide, such as weight,33 fatigue, and physical fitness,34 that we did not measure during the study. It is possible that any change in CC quality may not be secondary to fatigue. We did not mandate when rescuers had to switch, nor did we record when providers swapped, and so each team would have variation in duration of compressions provided by each rescuer (which may be more consistent with the real-world environment).


In a 12-minute simulated cardiac arrest, we demonstrated no decrease in quality of CC provided, although overall quality remained poor. More work is required to determine if JIT training and/or visual feedback may reduce the effect of rescuer fatigue in single-rescuer CPR.


The principal investigator, Jonathan Duff, and Yiqun Lin, MD, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Yiqun Lin (University of Calgary) conducted and is responsible for the data analysis.

INSPIRE group authors: Jennifer Davidson, RN, Alberta Children's Hospital, [email protected]; Dawn Taylor Peterson, PhD, Children's of Alabama, [email protected]; Marjorie Lee White, MD, MPPM, Med, Children's of Alabama, [email protected]; John Zhong, MD, Children's Medical Center of Dallas, [email protected]; Vincent Grant, MD, Alberta Children's Hospital, [email protected]; David Grant, MBChB, MRCPCH, Bristol Royal Hospital for Children, [email protected]; Stephanie Sudikoff, MD, Yale-New Haven Health, [email protected]; Kimberly Marohn, MD, Baystate Children's Hospital, [email protected]; Nicola Peiris, BSc, Alberta Children's Hospital, [email protected]; Jordan Duval-Arnould, MPH, CPH, DrPH[c], Johns Hopkins University School of Medicine, [email protected]; Ronald Gottesman, MD, Montreal Children's Hospital, [email protected]; Hubert Wong, MD, University of British Columbia, [email protected]; Aaron Donoghue, MD, MSCE, The Children's Hospital of Philadelphia, [email protected]; Robert M. Sutton, MD, MSCE, The Children's Hospital of Philadelphia, [email protected]; Dana Niles, MS, DPhil (candidate), The Children's Hospital of Philadelphia, [email protected]; Jenny Chatfield, Alberta Children's Hospital, [email protected]; Nnenna Chime, MD, MPH, Albert Einstein College of Medicine and Children's Hospital at Montefiore, [email protected]; Mark Adler, MD, MEd, [email protected]; Quhyn Doan, MD, [email protected].


1. Vaillancourt C, Everson-Stewart S, Christenson J, et al. The impact of increased chest compression fraction on return of spontaneous circulation for out-of-hospital cardiac arrest patients not in ventricular fibrillation. Resuscitation. 2011;82:1501–1507.
2. Sutton RM, French B, Niles DE, et al. 2010 American Heart Association recommended compression depths during pediatric in-hospital resuscitations are associated with survival. Resuscitation. 2014;85:1179–1184.
3. Idris AH, Guffey D, Aufderheide TP, et al. Relationship between chest compression rates and outcomes from cardiac arrest. Circulation. 2012;125:3004–3012.
4. Edelson DP, Abella BS, Kramer-Johansen J, et al. Effects of compression depth and pre-shock pauses predict defibrillation failure during cardiac arrest. Resuscitation. 2006;71:137–145.
5. Cheskes S, Schmicker RH, Christenson J, et al. Perishock pause: an independent predictor of survival from out-of-hospital shockable cardiac arrest. Circulation. 2011;124:58–66.
6. Christenson J, Andrusiek D, Everson-Stewart S, et al. Chest compression fraction determines survival in patients with out-of-hospital ventricular fibrillation. Circulation. 2009;120:1241–1247.
7. Atkins DL, Berger S, Duff JP, et al. Part 11: Pediatric basic life support and cardiopulmonary resuscitation quality: 2015 American Heart Association Guidelines Update for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation. 2015;132:S519–S525.
8. Abella BS, Alvarado JP, Myklebust H, et al. Quality of cardiopulmonary resuscitation during in-hospital cardiac arrest. JAMA. 2005;293:305–310.
9. Sutton RM, Niles D, French B, et al. First quantitative analysis of cardiopulmonary resuscitation quality during in-hospital cardiac arrests of young children. Resuscitation. 2014;85:70–74.
10. Arshid M, Lo TY, Reynolds F. Quality of cardio-pulmonary resuscitation (CPR) during paediatric resuscitation training: time to stop the blind leading the blind. Resuscitation. 2009;80:558–560.
11. Ashton A, McCluskey A, Gwinnutt CL, et al. Effect of rescuer fatigue on performance of continuous external chest compressions over 3 min. Resuscitation. 2002;55:151–155.
12. Bjorshol CA, Soreide E, Torsteinbo TH, et al. Quality of chest compressions during 10min of single-rescuer basic life support with different compression: ventilation ratios in a manikin model. Resuscitation. 2008;77:95–100.
13. Chi CH, Tsou JY, Su FC. Effects of compression-to-ventilation ratio on compression force and rescuer fatigue during cardiopulmonary resuscitation. Am J Emerg Med. 2010;28:1016–1023.
14. Haque IU, Udassi JP, Udassi S, et al. Chest compression quality and rescuer fatigue with increased compression to ventilation ratio during single rescuer pediatric CPR. Resuscitation. 2008;79:82–89.
15. Heidenreich JW, Berg RA, Higdon TA, et al. Rescuer fatigue: standard versus continuous chest-compression cardiopulmonary resuscitation. Acad Emerg Med. 2006;13:1020–1026.
16. Hightower D, Thomas SH, Stone CK, et al. Decay in quality of closed-chest compressions over time. Ann Emerg Med. 1995;26:300–303.
17. Jo CH, Cho GC, Ahn JH, et al. Rescuer-limited cardiopulmonary resuscitation as an alternative to 2-min switched CPR in the setting of inhospital cardiac arrest: a randomised cross-over study. Emerg Med J. 2014.
18. Manders S, Geijsel FE. Alternating providers during continuous chest compressions for cardiac arrest: every minute or every two minutes? Resuscitation. 2009;80:1015–1018.
19. McDonald CH, Heggie J, Jones CM, et al. Rescuer fatigue under the 2010 ERC guidelines, and its effect on cardiopulmonary resuscitation (CPR) performance. Emerg Med J. 2013;30:623–627.
20. Ochoa FJ, Ramalle-Gomara E, Lisa V, et al. The effect of rescuer fatigue on the quality of chest compressions. Resuscitation. 1998;37:149–152.
21. Badaki-Makun O, Nadel F, Donoghue A, et al. Chest compression quality over time in pediatric resuscitations. Pediatrics. 2013;131:e797–e804.
22. Li ES, Cheung PY, O'Reilly M, et al. Rescuer fatigue during simulated neonatal cardiopulmonary resuscitation. J Perinatol. 2015;35:142–145.
23. Cheng A, Brown LL, Duff JP, et al. Improving cardiopulmonary resuscitation with a CPR feedback device and refresher simulations (CPR CARES study): a randomized clinical trial. JAMA Pediatr. 2015;169:137–144.
24. Cheng A, Auerbach M, Hunt EA, et al. Designing and conducting simulation-based research. Pediatrics. 2014;133:1091–1101.
25. Sutton RM, Niles D, Meaney P, et al. "Booster" training: evaluation of instructor-led bedside cardiopulmonary resuscitation skill training and automated corrective feedback to improve cardiopulmonary resuscitation compliance of pediatric basic life support providers during simulated cardiac arrest. Pediatr Crit Care Med. 2011;12:e116–121.
26. Sutton RM, Maltese MR, Niles D, et al. Quantitative analysis of chest compression interruptions during in-hospital resuscitation of older children and adolescents. Resuscitation. 2009;80:1259–1263.
27. Rittenberger JC, Guimond G, Platt TE, et al. Quality of BLS decreases with increasing resuscitation complexity. Resuscitation. 2006;68:365–369.
28. Bjorshol CA, Sunde K, Myklebust H, et al. Decay in chest compression quality due to fatigue is rare during prolonged advanced life support in a manikin model. Scand J Trauma Resusc Emerg Med. 2011;19:46.
29. Chung TN, Kim SW, You JS, et al. The specific effect of metronome guidance on the quality of one-person cardiopulmonary resuscitation and rescuer fatigue. J Emerg Med. 2012;43:1049–1054.
30. Sugerman NT, Edelson DP, Leary M, et al. Rescuer fatigue during actual in-hospital cardiopulmonary resuscitation with audiovisual feedback: a prospective multicenter study. Resuscitation. 2009;80:981–984.
31. Berden HJ, Bierens JJ, Willems FF, et al. Resuscitation skills of lay public after recent training. Ann Emerg Med. 1994;23:1003–1008.
32. Bhanji F, Donoghue AJ, Wolff MS, et al. Part 14: education: 2015 American Heart Association guidelines update for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation. 2015;132:S561–S573.
33. Hasegawa T, Daikoku R, Saito S, et al. Relationship between weight of rescuer and quality of chest compression during cardiopulmonary resuscitation. J Physiol Anthropol. 2014;33:16.
34. Ock SM, Kim YM, Chung J, et al. Influence of physical fitness on the performance of 5-minute continuous chest compression. Eur J Emerg Med. 2011;18:251–256.

cardiopulmonary resuscitation; fatigue; just-in-time training; feedback; simulation

Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.