A comparative study of artificial intelligence and human doctors for the\n purpose of triage and diagnosis

Abstract

Online symptom checkers have significant potential to improve patient care,\nhowever their reliability and accuracy remain variable. We hypothesised that an\nartificial intelligence (AI) powered triage and diagnostic system would compare\nfavourably with human doctors with respect to triage and diagnostic accuracy.\nWe performed a prospective validation study of the accuracy and safety of an AI\npowered triage and diagnostic system. Identical cases were evaluated by both an\nAI system and human doctors. Differential diagnoses and triage outcomes were\nevaluated by an independent judge, who was blinded from knowing the source (AI\nsystem or human doctor) of the outcomes. Independently of these cases,\nvignettes from publicly available resources were also assessed to provide a\nbenchmark to previous studies and the diagnostic component of the MRCGP exam.\nOverall we found that the Babylon AI powered Triage and Diagnostic System was\nable to identify the condition modelled by a clinical vignette with accuracy\ncomparable to human doctors (in terms of precision and recall). In addition, we\nfound that the triage advice recommended by the AI System was, on average,\nsafer than that of human doctors, when compared to the ranges of acceptable\ntriage provided by independent expert judges, with only a minimal reduction in\nappropriateness.\n

Ähnliche Arbeiten