The Trip Report: Measuring Bystander CPR Quality

Reviewed This Month

Analysis of Bystander CPR Quality During Out-of-Hospital Cardiac Arrest Using Data Derived From Automated External Defibrillators.  

Authors: Fernando SM, Vaillancourt C, Morrow S, Stiell IG.

Published in: Resuscitation, 2018 May 16; 128: 138–43.  

The importance of high-quality CPR has been ingrained in our minds for years. We train often to maintain and improve our skills because high-quality CPR leads to improved survival for our patients.

Bystander CPR and AED use are also associated with improved survival rates, and bystander CPR has been shown to have a positive association with better neurologic outcomes for out-of-hospital cardiac arrest. However, few studies have examined the quality of CPR provided by bystanders. The limited research is likely due to the difficulty in collecting bystander CPR quality data. Luckily for us, technology is making it easier. 

Canadian emergency physician Shannon Fernando, MD, et al., recently published a study examining bystander CPR quality during out-of-hospital cardiac arrest using data derived from AEDs. To do this, the authors utilized data collected by the Resuscitation Outcomes Consortium (ROC) cardiac arrest database. The ROC is a network of clinical and satellite centers that collaborate to facilitate clinical trials and other outcomes-based cardiopulmonary arrest and severe traumatic injury research. The ROC database used for this study was a deidentified registry of calls in which EMS cared for a patient suffering from out-of-hospital cardiac arrest. To be included in this database, the OHCA patient had to be evaluated by EMS and have had attempts at external defibrillation by a layperson or emergency provider. Cases were also included if CPR was provided by healthcare professionals who were not 9-1-1 response personnel. Finally, cases were included if the patient was pulseless but did not have defibrillation attempted or CPR provided by EMS.

Patients were excluded if they were pregnant; prisoners; had a DNR; had a blunt, penetrating, or burn-related injury; or had a severe loss of blood. Authors only analyzed patients who were at least 18. They also only included patients whose OHCA was of presumed cardiac etiology. The study period was from July 2011 to July 2016.

The ROC database data was matched with data obtained from 550 AED uses from the Ottawa Paramedic Service’s AED database. Matched cases had a date and time within 4 minutes. All the AEDs were located in public buildings in the Ottawa-Carleton region, and they all had the ability to measure the quality of CPR using an accelerometer between the rescuer and patient’s chest. The AEDs also provided audio CPR quality feedback to the rescuer. The authors analyzed the first five minutes of each AED use. 

They also used the Utstein definition of a bystander, which is “any person involved in a resuscitation who is not responding as part of an organized emergency response system to a cardiac arrest.” To ensure CPR quality could accurately be assessed, cases were only included if they had at least a minute’s worth of CPR data recorded. The authors decided this inclusion criteria was necessary because those cases with shorter durations often represented shockable rhythms and were more likely to achieve return of spontaneous circulation with improved survival. 

The outcome measures used to evaluate CPR quality included the chest compression fraction (the proportion of time spent providing chest compressions during resuscitation efforts), average compression rate, average compression depth, and average perishock pause. The perishock pause was the sum of the preshock pause (the time between stopping chest compressions and delivering a shock) and postshock pause (the time from shock delivery to resuming chest compressions). The authors measured adherence to either the 2010 or 2015 American Heart Association guidelines as a secondary outcome of interest. 

A Very Small Percentage

During the study period there were 4,274 OHCAs in the Ottawa-Carleton region. The authors removed 4,126 (97%) cases from the study analysis because these cases met at least one exclusion criteria. The largest percentage (45%) were excluded because they were nontreated. An example of a nontreated case would be a patient with a DNR. 

The second-largest percentage (39%) of cases excluded were those where no AED was applied or used. The remainder were removed because either the arrest was of a presumed noncardiac etiology, EMS witnessed the arrest, there was no available CPR quality data recorded by the AED, or less than one minute of CPR data was available. 

The final study population included 100 cases. It’s easy to understand why so many cases were excluded. The inclusion and exclusion criteria likely led to the most appropriate study population. However, the authors ended up with a very small percentage of the overall population for analysis. Technology has made collecting this type of data easier, but it’s still hard to collect data to evaluate bystander CPR quality. 

Results

The authors did not find any meaningful differences in patient characteristics between the study population and those who could not be included due to missing data. There were 79 cases in which the AED was used by a police officer and 21 where the AED was used by the general public. There were no statistically significant differences noted when comparing the CPR quality of police officers and the general public (all p-values > 0.05). Overall, the average chest compression fraction was 76%, while the average depth was 5.3 cm and average compression rate was 111.2 per minute. 

Bystanders performed better when basing the assessment on the 2010 guidelines vs. the 2015 guidelines. When evaluating adherence to the 2010 guidelines, 66% adhered to the recommended rate of greater than 100 compressions per minute, 55% adhered to the recommended depth of greater than 5.0 cm, and 41% adhered to both the recommended rate and depth. When assessing adherence to the 2015 guidelines, 50% adhered to the recommended compression rate of between 100–120/min., 29% adhered to the recommended compression depth of between 5–6 cm, and 16% adhered to both the rate and depth recommendations. Average perishock pause was about 27 seconds. 

Interestingly, the authors found that for all measures of CPR quality as well as adherence to the 2010 and 2015 guidelines, performance was worse in the first minute of the resuscitation. This improved across the board in the second minute and remained consistent throughout the final four minutes of the evaluation period. The authors suggested this could be due to the programming of the AEDs: They noted the average time from turning on the AED to analyzing the rhythm was over 40 seconds, and the rhythm analysis took almost 10 seconds.

Due to this finding, the authors recommended that AED manufacturers program their machines to prompt bystanders to immediately begin chest compressions and that artifact filtering technology be utilized to allow for rhythm analysis during compressions. They further postulated that deficiencies in rate and depth of CPR could be due to difficulty estimating measurements and fear of causing harm. They noted this is likely alleviated after feedback is provided. 

Limitations

Like every study, there are limitations here, and the authors describe them well. The most glaring is the exclusion of about 98% of the population of OHCA patients. This could introduce selection bias. Also, results may not be generalizable, particularly when compared to performance when CPR feedback is not provided. Finally, there is no way to know the level of training or how many bystanders participated in the resuscitation.

This was an interesting study with well-thought-out case definitions and appropriate statistical analyses. it found that bystanders who step up to intervene in OHCA can and, more often than not, do provide quality CPR. If you haven’t found a way to promote bystander CPR in your area, hopefully this study provides motivation. 

Antonio R. Fernandez, PhD, NRP, FAHA, is research director at the EMS Performance Improvement Center and an assistant professor in the Department of Emergency Medicine at the University of North Carolina–Chapel Hill.