Introduction: High-quality cardiopulmonary resuscitation (CPR) is a key factor to cardiac arrest survival. Accurate monitoring and real-time feedback are emphasized to improve CPR quality. The purpose of this study was to develop and validate a novel depth estimation algorithm based on a smartwatch with built-in accelerometer for feedback instruction during CPR.
Methods: For data collection and model building, an Android Wear was worn by a researcher performing compression-only CPR on Resusci Anne manikin with QCPR sets (as reference standard). Acceleration data was collected by the 3-axis accelerometer of the smartwatch and processed to eliminate gravitational influence. The corresponding values of chest compression depth (CCD) were labeled according to the reference. We developed an algorithm based on the assumption that 1) maximal acceleration measured by the smartwatch accelerometer and the CCD are positively correlated and 2) the magnitude of acceleration at a specific time point and interval is correlated with its neighboring point. We defined a statistic value M as a function of time and the magnitude of maximal acceleration. Data were collected and processed and the relationship of M value, chest compression rate (CCR) and CCD were determined. A smartwatch app capable of detecting CCD was developed accordingly. During the validation process, a 2nd researcher wearing a smartwatch with pre-installed app performed compression-only CPR on the manikin at target sessions. The CCD results given by the smart watch and the reference were compared using the Wilcoxon Signed Rank Test (WSRT). Bland-Altman (BA) analysis was used to assess the agreement between two methods.
Results: For target CCR of 100-120/min and CCD of 5-6 cm, 2159 total compressions were analyzed. WSRT showed that there was no significant difference between the 2 methods (P=0.084). By BA analysis the mean of differences was 0.0135 and the bias between 2 methods was not significant (95% CI -0.0003 to 0.0274).
Conclusions: Our preliminary study indicates that the algorithm developed for estimating CCD based on a smartwatch with built-in accelerometer is promising. Further studies will be conducted to validate a broader range of CCR and CCD, and its application for clinical CPR training.