Investigating BLS instructors’ ability to evaluate CPR performance: focus on compression depth, rate, and recoil

Background: Out-of-hospital cardiac arrest (OHCA) presents significant challenges with low survival rates, emphasizing the need for effective bystander CPR training. In Basic Life Support (BLS) training, the role of instructors is pivotal as they assess and correct learners’ cardiopulmonary resuscitation (CPR) techniques to ensure proficiency in life-saving skills. This study evaluates the concordance between CPR quality assessments by Basic Life Support (BLS) instructors and those determined through Quantitative CPR (QCPR) devices, utilizing data from BLS courses conducted at National Cheng Kung University Hospital from October 2017 to April 2018. Methods: The study analyzed existing data from BLS courses, comparing CPR quality assessments made by instructors with those recorded by QCPR devices. Key metrics such as chest compression speed, depth, and recoil were examined to identify the degree of consistency between human and automated evaluations. Results: In this study, CPR performance was analyzed using QCPR devices and BLS instructors across metrics like speed, depth, and recoil. Employing the Cohen kappa statistic revealed moderate to low interrater reliability, the kappa value is 0.65 (95% C.I. 0.65–0.65) for depth, 0.56 (95% C.I. 0.33–0.79) for speed, and 0.50 (95% C.I.0.28–0.71) for recoil. Correlation analysis visualized in a heatmap indicated a higher consistency in depth evaluations (correlation coefficient = 0.7) compared to speed and recoil, suggesting a need for improved alignment in CPR training assessments. Conclusions: The study underscores the importance of refining CPR training methods and adopting advanced technological aids to enhance the reliability of CPR skill assessments. By improving the accuracy of these evaluations, the training can be better tailored to increase the effectiveness of life-saving interventions, potentially boosting survival rates in out-of-hospital cardiac arrest scenarios.