Abstract
Cardiac arrest is a life-threatening emergency that requires effective cardiopulmonary resuscitation (CPR) to preserve life. Smartwatches equipped with accelerometer sensors and worn close to the wrist joint may offer an accessible way to provide CPR feedback. This study determined whether accelerometer data from a smartwatch could provide valid estimates of CPR parameters, including compression, decompression, frequency, and cycle time. Nineteen medical school students performed CPR on an instrumented mannequin while wearing a smartwatch. CPR parameters were derived from the smartwatch’s accelerometer signals using a custom algorithm after first resampling the data to 60 Hz to circumvent the issue of small fluctuations in the sampling frequency inherent in these devices. Results were compared with kinematic data obtained from video analysis undertaken simultaneously. Across 3,005 CPR cycles, smartwatch analysis indicated that 76.7% of compressions were within the recommended depth range (5–6 ± 0.5 cm), while the mannequin software reported a mean performance score of 98.7 ± 1.35%. There was good agreement between smartwatch and video analysis for cycle duration (bias = 0.00 s; 95% CI: −0.05 to 0.05 s) and compression depth (bias = − 0.31 cm; 95% CI: −1.61 to 0.98 cm). These findings demonstrate that smartwatch accelerometer data analyzed with the proposed algorithm can accurately assess CPR performance administered to a mannequin, supporting its potential use as a practical assessment tool for CPR. Further work is needed to develop a smartwatch App that provides real-time feedback to improve CPR performance during training and real-world resuscitation scenarios. GRAPHICAL ABSTRACT: [Image: see text] SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s11517-026-03533-z.