Benefit-Risk Reporting for FDA-Cleared Artificial Intelligence− Enabled Medical Devices

This cross-sectional study examined how comprehensively the FDA-cleared artificial intelligence and machine learning medical devices report efficacy, safety, and risk data, analyzing all 691 such devices cleared from September 1995 to July 2023. Key reporting elements were frequently absent: study design was unreported for nearly half of devices, training sample size for over half, and demographic representation for 95.5% of devices. Only six devices reported data from randomized clinical trials, fewer than three reported patient outcomes, premarket safety assessments were documented for just 28.2% of devices, and postmarket adverse events were reported for only 5.2%.

These findings expose a significant gap between the rapid proliferation of AI-enabled medical devices and the regulatory infrastructure designed to ensure their safety. The study underscores an urgent need for dedicated regulatory pathways and more robust postmarket surveillance systems, as the current FDA clearance process allows the vast majority of AI devices to enter clinical use without rigorous clinical testing or standardized reporting of their real-world performance and risks.