We conducted a prospective, randomized controlled trial with 2 parallel groups of voluntary pediatric residents comparing time to first defibrillation attempt while using AR glasses (Google Glass, allocation group A) or AHA PALS conventional pocket reference cards (allocation group B) during a standardized simulation-based pediatric CA scenario (Multimedia Appendix 1). No changes were made to the AR glasses or the intervention during the study.

In March 2016, 20 pediatric residents participated and completed the study with no dropout (Figure 3). The demographic results are summarized in Table 1.

The first defibrillation was delivered within 180 seconds by 70% of the residents in both groups (Figure 4). Mean times to resuscitation critical endpoints are summarized in Table 2. None of them were significantly different between allocation groups A and B. All participants (100%) correctly recognized the pVT rhythm on the monitor. In both allocation groups, 90% of the residents initiated chest compressions within 60 seconds from the onset of pVT, and half of them before 20 seconds. There were no statistically significant differences in hands-on time spent by cycles of chest compression between both groups. Due to a lack of some defibrillation attempts, 1 interval in allocation group A and 4 in group B were not measurable.

At the time of the study, 6 participants (60%) in allocation group A and 7 participants (70%) in allocation group B were PALS-certified providers. Eleven participants out of 20 (55%) were residents with more than 1 year of PGY. With regard to all outcomes measured, we observed in both allocation groups that PALS-certified residents or those with PGY>1 tended to defibrillate and deliver drugs more quickly than non-PALS residents or those with ≤1 PGY (Multimedia Appendix 4). We observed no difference with previous resuscitation experience (data not shown).

Errors and deviations from the AHA pVT algorithm are summarized in Table 3. Out of 40 opportunities, 5 errors in defibrillation doses (12.5%) were committed during the whole scenario in allocation group A. This compares to 26 errors in defibrillation doses (65%) during the whole scenario in allocation group B (P<.001). Out of 10, one resident (10%) in allocation group B wrongly used a synchronized shock, 2/10 (20%) never delivered a third defibrillation, and 2/10 (20%) stopped the compressions after the 1st or 2nd defibrillation attempts.

The entire pVT algorithm was followed correctly in a stepwise fashion until ROSC by 60% of residents in allocation group A and 40% in allocation group B (P=.66).

The questionnaire was completed and returned by 100% of the participants. Participants in allocation group A and B rated the overall perceived stress to be 6.2 (95% CI 4.7-7.7) and 7.0 (95% CI 5.7-8.3), respectively on the Likert scale (P=.38). The usability, acceptance, and perception of the AR glasses were a major concern in our study and were assessed using a 17-item Unified Theory of Acceptance and Use of Technology (UTAUT) questionnaire [34]. The results will be published in a separate upcoming study.

To our knowledge, this is the first randomized controlled trial to investigate the benefit of a wearable technology to improve pediatric residents’ performance and adherence with regard to AHA resuscitation guidelines. Using AR glasses, we found that time to first defibrillation attempt, time to other critical resuscitation endpoints, and drug dose delivery were not improved in terms of adherence to AHA guidelines. However, errors and deviations from the pVT guideline in terms of defibrillation doses and cumulative defibrillation doses were significantly reduced when compared with the use of the PALS pocket reference cards.

During resuscitation, time is a decisive success criterion. During the first 15 minutes, survival and favorable neurological outcome decrease linearly by 2.1% and 1.2% per minute, respectively [35]. Delays in initiating CPR have a detrimental effect on patient outcome regardless of the quality of resuscitation [36]. AHA therefore recommends pulseless patients of any age to receive immediate CPR without delay starting with chest compressions followed by a defibrillation within 180 seconds of a shockable rhythm. However, management, procedural skills, and adherence to these guidelines have been shown to fade after a few months of initial training [37-39]. With a critical patient’s condition and stress, physicians do not always have enough time to apply these guidelines and are prone to deviate from them [40]. PALS pocket cards are intended to resolve this problem by delivering fast and accurate summarized resuscitative knowledge and skills to providers. Nevertheless, Hunt et al have observed that despite availability of these recommendations, 66% of pediatric residents failed to start compressions within 60 seconds from the onset of a simulated pVT, 33% never started compressions, only 54% successfully defibrillated within 180 seconds, and 7% never discharged the defibrillator [40]. Similar results were obtained by Labrosse et al during a simulated pulseless shockable arrest scenario, where 25% of pediatric residents failed to start compressions and 4% never defibrillated a patient [41]. A more recent study among first-year pediatric residents showed a median time for initiation of CPR of 50 seconds and to first defibrillation of 282 seconds [42]. Pediatric residents performed better in our study with delays closer to AHA recommendations with a mean time to initiate chest compressions of 25 to 28 seconds and to first defibrillation of 146 seconds. In both allocation groups, 90% of residents started compressions within 60 seconds from the onset of pVT, and 70% defibrillated within 180 seconds. However, there were no advantages for residents to wear the AR glasses as they performed similarly with or without them regarding delay to critical resuscitation endpoints, whether they were PALS-certified or not. Despite an ergonomic-driven approach to adapt AHA resuscitation algorithms in AR glasses and a prior 15-minute training session for their use, our system failed to improve resuscitation efficiency in terms of time to major endpoints. An explanation might be that reducing further time to defibrillations and drug delivery was not achievable by residents training in emergency medicine. Indeed, our results in accordance with those from Hunt et al [40] showed a trend toward improvement over PGY of training and PALS certification in the mean time to all defibrillation attempts. It would be interesting in further studies to assess this assumption with certified emergency physicians.

Current AHA resuscitation guidelines emphasize 2 minutes of chest compressions between defibrillations attempts as optimal care for persistent pVT or VF in children [30,43]. In this study, residents performed similarly on average, with or without AR glasses and close to AHA recommendations.

Prompt defibrillation is crucial for termination of VF or pVT to achieve ROSC [43]. The AHA 2015 guidelines recommend treating pVT or VF in children with an initial dose of 2 J/kg [30]. For subsequent shocks, a dose of 4 J/kg is recommended, though higher energy levels may be considered up to adult dose, if not exceeding 10 J/kg (Figure 1). In this trial, residents using the PALS pocket cards were more prone to deviate from defibrillation doses than those using the AR glasses. On average, the shocks they provided were delivered with defibrillation doses 13% to 73% above AHA recommendations. “High” defibrillation doses concerned mostly the initial shock, with doses reaching up to 6 J/kg in 40% of cases when delivered by residents not wearing the glasses. In an observational study of 285 pediatric IHCA, a higher initial shock dose of more than 2 J/kg was not associated with superior termination of pVT or VF or improved survival rates [44]. In addition, children who were defibrillated with higher initial shock doses in the >3-5 J/kg range were significantly less likely to have termination of pVT or VF with ROSC or to survive the event. In our study, the final cumulative defibrillation dose delivered by the residents in the PALS pocket cards group was on average 33% significantly higher than the AHA expected value. This deviation was 10 times greater than that seen with residents wearing the AR glasses. In particular, 50% of residents using the PALS pocket cards used wrong energy doses whether it was for first or second shock delivery. In a total of 40 defibrillation attempts, they used wrong doses in 65% of cases for a cumulative defibrillation dose of 18.7 J/Kg. These errors were reduced by 53% and cumulative dose by 37% by using the AR glasses, suggesting a limited but worthwhile benefit of their use in simulated resuscitation. It would be interesting in further studies to determine whether this would translate into fewer errors in shock doses in real life.

Finally, in terms of drug dose concentrations, both groups in this study accurately administered epinephrine and amiodarone. The entire pVT algorithm was followed correctly until ROSC in a stepwise fashion by 60% of residents wearing AR glasses, compared with 40% of residents with the PALS pocket cards.

This study has some limitations. First, it was conducted in a simulation-based resuscitation scenario. This choice was related to the ethical and organizational difficulties of conducting studies in real-life critical conditions. However, several studies have demonstrated the benefit of simulation as an investigative research methodology to answer research questions that otherwise could not be answered during CPR [45]. High-fidelity simulation is recognized as an essential tool to study resuscitation skills or technologies. Till date, none of the results obtained from simulation-based CPR studies disagreed with those obtained from studies in real life, confirming our study design choice. Realism was achieved, reflected by the stress levels experienced by the participants. They quoted the simulation as highly stressful when compared with real CPR situations.

Second, one might raise questions of our choice to choose Google Glass as the best AR glasses to display AHA guidelines. Indeed, despite its remarked 2012 commercial unveiling, Google Glass never reached its public audience and was discontinued in January 2015. However, there is a growing interest in recent literature toward its use in specialized medical fields [21,22,25,46,47]. We hypothesize that other AR glasses would not drastically change the results that we found with Google Glass, as their small size remains a major limitation to their use in displaying CPR algorithms. Further studies would be valuable to assess this assumption.

Taken together, our results support the interpretation that residents are not accurately following AHA recommendations during pediatric resuscitation, whether they are PALS certified or not. A wearable technology such as AR glasses might partially fill this gap and benefit patients by improving adherence and performance of residents to meet resuscitation requirements set by AHA, especially regarding delivery of defibrillation doses. In this sense, AR glasses appear as an interesting tool for emergency medicine and future studies are required to further examine this new paradigm.