Machine learning can support dispatchers to better and faster recognize out-of-hospital cardiac arrest during emergency calls: A retrospective study

AimFast recognition of out-of-hospital cardiac arrest (OHCA) by dispatchers might increase survival. The aim this observational study emergency calls was to (1) examine whether a machine learning framework (ML) can the proportion recognizing OHCA within first minute compared with dispatchers, (2) present performance ML different false positive rate (FPR) settings, (3) call characteristics influencing recognition.MethodsML be configured FPR i.e., more or less inclined suspect an depending on predefined setting. is evaluated 1.5 as primary endpoint, and other settings secondary endpoints. exposed random sample from 2018. Voice logs were manually audited evaluate time recognition.ResultsOf 851 calls, recognized 36% (n = 305) 1 min 25% 213) dispatchers. at any during 86% for 84% median 72 versus 94 s. both dispatcher showed 28 s mean difference in favour (P < 0.001). higher reduced times.ConclusionML has potential supportive tool calls. optimal need prospective study.