Impact of a Quality Cardiopulmonary Resuscitation Coach on Pediatric Intensive Care Unit Resuscitation Teams

Many children who sustain an in-hospital cardiopulmonary arrest do not survive.^1,2  When resuscitation teams rely solely on visual assessment to determine the quality of cardiopulmonary resuscitation (CPR), they often overestimate chest compression rate, depth, and proportion of time in the target range.³  Emerging evidence has shown that adherence to American Heart Association (AHA) CPR quality metrics improved during simulated arrests when resuscitation teams incorporated a feedback device and a quality CPR coach.^4–6 

The CPR metrics recommended by the AHA are based on evidence that supports survival and favorable neurologic outcomes.^1,7,8  Despite experience and training, resuscitation teams across institutions vary in their adherence to CPR metrics and the quality of CPR provided.^9–17 

The role of a quality CPR coach was developed for hospital-based resuscitation teams. The coach is a supplementary team member who provides real-time, verbal feedback on chest compression quality to compressors during a cardiac arrest.⁶  The goal is to decrease cognitive workload for the team leader, allowing the leader to focus on advanced life support, including identifying and managing reversible causes of the arrest.^6,18,19  The AHA offers training courses to teach CPR and other lifesaving management and principles; however, these courses do not yet incorporate targeted training on the CPR coach role. Efforts are ongoing to standardize and disseminate CPR coach competency training, but this training is not yet widely available.^20,21 

This study was conducted at an academic, free-standing children’s hospital in the midwestern United States with a 72-bed pediatric and cardiac intensive care unit (PICU). The PICU averaged 37 in-unit compression events per year from 2018 to 2021. In 2017, the institution switched defibrillator devices to the Zoll defibrillator. The Zoll defibrillator functions as both a defibrillator and a feedback device to provide real-time visual information on a display regarding the depth and rate of chest compressions using Real CPR Help technology (Zoll Medical Corporation). The feedback function requires that adhesive pads be placed on the patient’s chest and back during compressions at the onset of the arrest. This pad placement and timing contrast with standard defibrillator devices that lack a feedback function, for which pads are needed only if defibrillation is required.

In addition to the transition to the Zoll defibrillator, the institution incorporated a CPR coach during compression events as part of a resuscitation quality initiative bundle. Initially, there was minimal targeted training, and health care professionals were oriented to the coach role as events naturally occurred over time. This resulted in variable Zoll defibrillator pad application, suboptimal use of visual feedback data, and inconsistent use of a code team member in the coach role.

To address these issues, we designed, implemented, and evaluated a CPR coach training curriculum. The training consisted of a series of progressive simulation cases using the rapid cycle deliberate practice method. Rapid cycle deliberate practice is a simulation-based education method that focuses on rapid acquisition of necessary skills to improve performance in low-volume, high-risk patient events^22,23  and that has been used effectively for quality CPR coach training in other studies.^4,6 

We defined short-term learning outcomes by using an educational planning and evaluation framework, the Kirkpatrick levels of evaluation.²⁴  The levels represent a hierarchy from low- to high-level outcomes as follows: level 1, reactions; level 2, knowledge; level 3, behavior; and level 4, results.²⁴  We implemented the curriculum in a series of educational sessions to train the critical care pediatric nurse practitioners. We report the educational intervention details and learning outcomes separately.²⁵ 

The purpose of this study was to evaluate the impact of a quality CPR coach training educational intervention and implementation of trained CPR coaches on resuscitation teams in a single center’s PICU. We used a logic model to clearly portray the connections between the educational intervention and patient-level outcomes. A logic model is a systematic and visual approach to presenting relationships among resources, activities, targeted changes, and results. A basic logic model includes components that describe the planned work (inputs and activities) and intended results (outputs, outcomes, and impact) (Figure 1).²⁶  In this article, we report the Kirkpatrick level 4 (results) outcomes, or the degree to which targeted patient-level outcomes occurred associated with the training,²⁴  including physiologic metrics of the quality of CPR delivery during compression events in a pediatric cardiac arrest cohort.

This project was determined to be exempt by our organization’s human research protection program office. The short-term outcomes, long-term outcomes, and impact metrics of the quality CPR coach intervention are described in Figure 1. This study primarily addresses the long-term outcomes, which align with Kirkpatrick level 4 (results) and include increased presence of a CPR coach on resuscitation teams, increased application of Zoll defibrillator adhesive pads, and improved quality of CPR delivery in real code events according to AHA-defined quality metrics.

We conducted a retrospective medical record review of patients who sustained in-unit cardiopulmonary arrests. Data collection included internal and external physiologic metrics of quality CPR delivery. Internal metrics were directly measured from the patient and included arterial diastolic blood pressure (aDBP) and end-tidal carbon dioxide capnography.⁶  Arterial DBP is a surrogate measure of aortic end-diastolic pressure to determine coronary perfusion pressure.^27–29  End-tidal carbon dioxide capnography is a surrogate measure of pulmonary blood flow and cardiac output.^30,31  External metrics were displayed or calculated by the Zoll defibrillator and included chest compression rate, chest compression depth, and chest compression fraction (CCF).⁶  The CCF is the proportion of total CPR time that chest compressions are performed during an arrest and is reported as a percentage. Reducing the number and duration of pauses and ensuring a compression rate of 100 to 120 beats per minute yields a higher CCF.^1,7,28,29  Internal metrics were captured on bedside cardiac monitors (GE Healthcare) and stored through BedMaster Solutions (Anandic Medical Systems). External metrics were captured via the Zoll defibrillator and were stored through RescueNet (Zoll Data Systems).

Compression event data are routinely collected and analyzed independent of this study as part of the institution’s internal PICU code review process. The study team was provided compression event data by a critical care quality improvement specialist who serves as a leader of the PICU code review committee. Deidentified data were reported and analyzed as dichotomous yes or no responses as follows: (1) CPR coach present; (2) Zoll defibrillator pads placed on the patient; (3) CCF at or above 60%; (4) CCF at or above 80%; (5) end-tidal carbon dioxide capnography metric available; (6) end-tidal carbon dioxide capnography maintained at greater than 15 mm Hg; (7) aDBP metric available; and (8) aDBP maintained at greater than 25 mm Hg (for patients <1 year of age) or at greater than 30 mm Hg (for patients ≥1 year of age).

Justification and supporting evidence for the target quality metrics were informed by the AHA and other expert consensus sources. The AHA cites a CCF goal of at least 60%, whereas expert consensus cites at least 80%,^1,7,28,29  and an end-tidal carbon dioxide goal of greater than 15 mm Hg, although others have suggested greater than 20 mm Hg and as close to 25 mm Hg as preferred.^6,7,30,31  Finally, an aDBP goal of greater than 25 mm Hg in infants and greater than 30 mm Hg in children is associated with increased likelihood of return of spontaneous circulation, survival to discharge, and survival with a favorable neurologic outcome.^1,27 

Inclusion criteria were that the cardiopulmonary arrest occurred in the PICU during 2018 to 2021, in a patient up to 18 years old, and lasted longer than 1 minute. Exclusion criteria were that the cardiopulmonary arrest occurred outside the PICU, in a patient older than 18 years, lasted for less than 1 minute, and entailed use of the LUCAS Chest Compression System. Multiple arrests of the same patient were excluded. Arrests that occurred during the training intervention period (30 September to 5 December 2019) were also excluded. Cardiac arrests lasting less than 1 minute were excluded because the quality CPR literature cites a key goal as ensuring that the Zoll defibrillator is turned on and that pads are applied within 2 minutes.⁶  Multiple arrests for the same patient were excluded because they could not be treated as independent samples, which was a statistical assumption that had to be met for the planned analysis.

SPSS (version 28) was used for the analyses. Data metrics were analyzed using descriptive statistics and are reported as frequencies and percentages. The Fisher exact test was performed to compare metrics from before to after the training. A P value of less than .05 was considered statistically significant.

The PICU had 148 cardiopulmonary arrests during the calendar years of 2018 through 2021. After we removed excluded events, 40 pretraining and 39 posttraining cardiopulmonary arrests remained for analysis (Figure 2). The age of the patients did not differ significantly between the 2 periods.

The presence of monitoring modalities available at the time of each arrest determined whether that CPR quality data metric was available in real time. The presence of each monitoring modality is reported as a frequency, and the monitoring modalities available did not differ significantly from before to after the intervention (Table 1). The Zoll defibrillator pads had to be placed on the patient for the CCF data to be calculated, and the Zoll feedback device had to be connected to the pads and turned on to display and collect the compression rate and depth data. End-tidal carbon dioxide capnography via an endotracheal tube had to be in place for end-tidal carbon dioxide measurement. An arterial blood pressure waveform via an in situ arterial catheter had to be in place for aDBP measurement.

Each quality CPR data metric was compared from before to after the training intervention (Table 2). Frequencies were calculated by using a denominator of the number of arrests eligible and a numerator of the number of arrests achieved (number of arrests achieved/number of arrests eligible). Cardiac arrests eligible refers to arrests in which the necessary monitoring modality was available (eg, an endotracheal tube was in place with an end-tidal capnography device to measure end-tidal carbon dioxide). Arrests achieved refers to arrests in which the defined goal for that data metric was successfully achieved during the arrest (eg, end-tidal carbon dioxide was greater than 15 mm Hg). Of the 40 pretraining arrests, 12 arrests had an aDBP measurement available (ie, the number eligible), and of those, 4 achieved aDBP above the goal range (ie, the number achieved). This finding is compared with the posttraining findings, for which, of the 39 arrests, 14 had an aDBP measurement available, and of those, 10 achieved aDBP above the goal range. Goal aDBP indicative of high-quality CPR delivery increased from 33% before training to 71% after training (P = .11). Presence of a CPR coach statistically increased significantly from 35% to 72% (P = .002). Zoll defibrillator pad placement increased from 50% to 72%, with near statistical significance (P = .07).

Training critical care pediatric nurse practitioners to function as quality CPR coaches on code teams yielded a statistically significant improvement in the frequency of a coach being present on the resuscitation team during real cardiopulmonary arrests in the PICU. Additionally, the frequency of the presence of a quality CPR coach and the frequency of Zoll defibrillator pad placement after the intervention were identical at 72%, which implies that if a CPR coach was present then the Zoll defibrillator pads were placed on the patients. Although we noted no statistically significant improvements in physiologic metrics indicating AHA-compliant quality CPR delivery, these data suggest that insertion of a trained CPR coach onto resuscitation teams is not harmful.

Several published simulation-based studies have collectively provided theoretical applications of the quality CPR coach role, with assumed translation to the clinical setting during real patient events.^{4–6,32–34}  Cheng and colleagues⁴  showed in a simulation-based randomized controlled study that the presence of a CPR coach compared with no coach in simulated arrests resulted in absolute improvement in overall high-quality CPR. In a large, multisite prospective simulation study, Kessler et al³³  demonstrated a reduced duration of pauses in chest compressions during simulated arrests when code teams included a CPR coach. The simulation-based studies laid the foundation and showed progressive improvement in achieving AHA-compliant quality CPR metrics from 12.7% adherence with no feedback device, to 38% adherence with visual feedback through a feedback device, and up to 69.5% adherence with both a feedback device and a CPR coach.^4,5 

In a recent editorial, expert resuscitation researchers released a call to action to insert trained CPR coaches onto real resuscitation teams and to measure the real-life effects.³⁴  Connecting an educational intervention to patient-level outcomes is one of the biggest challenges facing educational researchers. Few educational studies measure and report at the highest levels of evaluation.^35–38  One systematic review identified no simulation-based educational studies that measured Kirkpatrick level 4 outcomes.³⁵  There are increasing calls on health professional educators to achieve and measure the highest learning outcome levels that translate to the patient’s bedside and organization. The T1 level of translational research measures outcomes in a controlled laboratory setting, T2 translational research measures outcome transfer to clinical settings where health care is delivered, and T3 translational research measures outcome transfer to patients and/or public health improvement as a result of the educational interventions.³⁸ 

As suggested by McGaghie, one approach to increase the quality and rigor of translational research in health professional education is to “stretch” research end points to measure outcomes at the level of the patient in health care delivery settings.^38(p3) The study reported here serves as an exemplar of how a logic model can be used to organize the design, implementation, and evaluation of an educational intervention. When used in tandem with the Kirkpatrick levels of evaluation, use of a logic model promotes consideration of higher-level outcomes from the outset of planning, identifies pragmatic issues enabling achievement of the outcomes, and allows multimodal data evaluation.

The findings of this study reinforce previously published reports showing that the presence of a quality CPR coach on code teams is associated with both enhanced adherence to and measurable improvements in AHA quality CPR metrics, which are associated with improved survival outcomes.^4,6  In their landmark prospective observational study, Hunt and colleagues⁶  described the development and refinement of the CPR coach role as part of a resuscitation quality bundle (CODE ACES²). During the 3-year implementation period, Hunt and colleagues demonstrated that the adjusted marginal probability of a chest compression epoch meeting criteria for “excellent CPR” (defined as >90% of compressions within AHA-compliant target rate, depth, and CCF) improved from 19.9% to 44.3% in real patient cardiac arrests.⁶ 

Beyond the results reported by Hunt et al⁶  regarding the development of the coach role, we identified no further literature describing real patient outcomes associated with the CPR coach role. Thus, the results of the present study represent a valuable first step in evaluating the insertion of trained CPR coaches onto real resuscitation teams in the PICU. Moreover, from an education research perspective, our study demonstrated the feasibility and practical application of simulation-based education with measurable effects at the level of the patient. The training intervention yielded a statistically significant increase in the presence of a CPR coach on code teams, suggesting that targeted training leads to providers assuming the role in real arrests. Although we found no statistically significant posttraining improvements in physiologic metrics indicating quality CPR delivery, the frequencies of AHA-compliant CPR were either unchanged or in many cases improved, indicating clear clinical significance.

Clinical significance refers to clinically relevant results, such as when research outcomes can be used to assess the effectiveness of an intervention or when findings yield improvements in patients’ quality of life or functionality.^39–43  The delivery of higher-quality CPR in even one arrest is clinically meaningful, because that one arrest equates to a child with a greater likelihood of survival and a favorable neurologic outcome. Moreover, the AHA has highlighted educational efficiency and local implementation as key determinants in the formula for survival after cardiac arrest.^44,45  The present educational intervention and evaluation study function as a blueprint for other centers to incorporate evidence-based educational strategies into practice, study the real-life effects, and ultimately positively affect patients’ resuscitation outcomes. Overall, after the quality CPR coach training intervention, this study demonstrated a statistically significant improvement in coach presence on code teams, which was associated with clinically significant improvement in some metrics of quality CPR delivery in real cardiopulmonary arrests.

The results of these short- and long-term outcome evaluations can inform further evaluation of impact-level outcomes, including patient survival and neurologic outcomes. Determining the impact of CPR coaches on code teams will require further work inclusive of a revised research design with a larger sample size and a case-controlled prospective design. The effects of skill decay are described throughout the literature, which suggests that short-duration, spaced training sessions are optimal for skill retention.^9,11,44,46  One randomized trial found monthly training to be more effective than trainings at 3-, 6-, or 12-month intervals.⁴⁶  Another randomized trial showed that the maximum instructor-to-participant ratio of 1:6 was most effective for detecting participants’ performance errors in CPR training.⁴⁷  Evidence suggests that a sustainability plan should include frequent training sessions in small groups to refresh skills and practice application. Finally, revisions to the quality CPR coach curriculum will need to coordinate with AHA updates to life support guidelines and incorporate new evidence regarding CPR coach outcomes, personalized CPR, and best practices across institutions.

Limitations of the present study include small sample sizes arising from a single center, a convenience sampling strategy that was unable to be case controlled, limited information on patient and code team demographics, the retrospective design and the reliability of the data variables abstracted from the patient record, and, last, the fact that other variables most likely affecting the data were not measured or controlled. Critical care nurse practitioners were the trained cohort but may not have exclusively functioned in the CPR coach role, and who functioned in the role was not documented at the time of data collection.

The delivery of high-quality CPR in adherence with AHA guidelines is associated with improved survival outcomes after cardiopulmonary arrest. Studies have shown that inserting a trained CPR coach onto code teams improves physiologic metrics indicating high-quality CPR delivery in simulated arrests, with assumed translation to the patient bedside. This study represents the first step in answering the call from experts to insert trained quality CPR coaches onto code teams and study the real-life effects. Moreover, this educational intervention and evaluation study can function as a blueprint for other centers to pragmatically actualize evidence-based educational strategies, study the real-life effects, and ultimately affect patients’ resuscitation outcomes. Overall, the quality CPR coach training intervention demonstrated a statistically significant increase in coach presence on code teams, and this increased presence was associated with clinically significant improvements in some metrics of quality CPR delivery in real arrests.

This research was performed at the Medical College of Wisconsin, Department of Pediatrics, Section of Critical Care, and Children’s Wisconsin. Katie McDermott thanks the director, instructors, and mentors of the Master of Education in the Health Professions program at Johns Hopkins University for their guidance.