Effect of Virtual Reality Cardiopulmonary Resuscitation Training With In-Scene Feedback on Chest Compression Depth Accuracy at 4 Weeks in Laypersons: Randomized Controlled Trial (Preprint)

BACKGROUND Bystander cardiopulmonary resuscitation (CPR) improves survival after out-of-hospital cardiac arrest, and effective training of laypersons is therefore important. Virtual reality (VR) can place trainees in an immersive emergency scene and display real-time feedback linked to the responses of a virtual patient, although current guidelines advise against VR for CPR skill training. OBJECTIVE This study aimed to compare VR-based CPR training that provides in-scene, patient-linked feedback with conventional manikin training that provides feedback on the same compression parameters, using chest compression quality measured 4 weeks after training in laypersons. METHODS In this single-center, parallel-group randomized controlled trial, healthy adults without CPR experience were randomized 1:1 to VR or instructor-led manikin training. Both groups practiced on sensor manikins that provided feedback on compression depth, rate, full chest recoil, and hand position. The VR group viewed the feedback within the virtual scene, and the virtual patient regained consciousness when compressions met the target range. Four weeks after training, a blinded assessor evaluated compression quality during a simulated cardiac arrest using a Resusci Anne QCPR manikin. The primary outcome was the proportion of compressions with a depth of 5–6 cm. Secondary outcomes included mean compression depth and rate, scenario response times, and satisfaction with training. RESULTS Of 99 adults screened, 94 were randomized (47 per group), and all completed the 4-week assessment. The proportion of compressions at 5–6 cm was 55.2% (SD 28.0%) in the VR group and 43.1% (SD 29.8%) in the conventional group (mean difference 12.1 percentage points, 95% CI 0.3-24.0; P=.045). Mean compression depth (55.1 vs 55.6 mm; P=.65), mean compression rate (117.3 vs 119.7/min; P=.21), full chest recoil, hand position, scenario response times, and satisfaction with training did not differ between groups. Two VR participants reported mild dizziness and one reported nausea without discontinuing training. No adverse symptoms occurred in the conventional group. CONCLUSIONS VR training with in-scene, patient-linked feedback improved the accuracy of chest compression depth at 4 weeks compared with conventional manikin training, without differences in mean depth, rate, or scenario response times. These findings support further evaluation of manikin-integrated VR for layperson CPR training. CLINICALTRIAL Clinical Research Information Service KCT0012772; (https://cris.nih.go.kr/cris/search/detailSearch.do?seq=35160) (registered retrospectively)