Effectiveness of rescue Me CPR! smartphone app providing real-time guidance to untrained bystanders performing CPR

Background: Out-of-hospital cardiac arrest (OHCA) is a persistent global health challenge, owing, in part, to low rates of population CPR training. Smartphone applications have the potential to widely disseminate CPR basic training to a populace, but other studies have found multiple limitations in previously developed CPR guidance applications (CPR-GA). This study aims to use medical simulation to assess the relative CPR performance of novices using the 'Rescue Me CPR!' (RMC) app, a custom CPR-GA designed by this research team, to novices using 'PG-CPR!' (PGC), the most downloaded CPR-GA available in the USA, and to CPR certified medical personnel. Methods: In a prospective randomized experimental trial of 60 individuals, subjects were either given the RMC app, the PGC app, or had active CPR certification. They were presented a cardiopulmonary arrest scenario and were observed while performing CPR on a high-fidelity manikin. Data was collected through four cycles of CPR, during which time 24 pertinent performance metrics and CPR steps were timed and recorded. These metrics were assessed on their own and used to calculate average time to compressions, average chest compression fraction, and rate of high-quality CPR for each study group. Results: CPR certified subjects called 911 in 100 % of simulation cases, started compressions 34 +/- 10 s after first seeing the simulated patient, had an average chest compression fraction of 0.52, and performed high-quality CPR in 25 % of aggregate compression cycles. PGC app users called 911 in 70 % of simulation cases, started compressions 86 +/- 17 s after first seeing the simulated patient, had an average chest compression fraction that could not be assessed due to inconsistent pauses during CPR, and performed high-quality CPR in 2.5 % of aggregate compression cycles. RMC app users called 911 in 100 % of simulation cases, started compressions 55 +/- 6 s after first seeing the simulated patient, had an average chest compression fraction of 0.48, and performed high-quality CPR in 50 % of aggregate compression cycles. Conclusion: The results of this study demonstrate that in all studied metrics, except time-to-firstcompression, CPR provided by individuals using the RMC app is statistically equivalent or superior to CPR performed by a CPR certified individual and, in almost every metric, superior to CPR performed by users of the most downloaded android CPR guidance application, PG-CPR.