The paper “Clinical Evidence and FDA Recalls of Artificial Intelligence–Enabled Medical Devices”, co-authored by Prof. Enrico Caiani and Dr Yijun Ren from the D-Hygea Lab, Department of Electronics, Information and Bioengineering at Politecnico di Milano, has been published in JAMA Network Open.
The study investigates the factors associated with the risk of recall for medical devices based on artificial intelligence, with the aim of improving post-market surveillance activities.
The research examined 903 devices authorized by the U.S. Food and Drug Administration, integrating the data with reports of adverse events and safety notices published in several EU and extra EU countries, obtained automatically by a digital platform previously developed at D-Hygea Lab.
The findings show a higher likelihood of recall for devices lacking documented clinical evidence. Misuse was also frequently associated with recalls, highlighting specific safety challenges related to artificial intelligence systems.
The study therefore emphasizes the importance of adopting an integrated surveillance framework capable of identifying risk signals at an early stage and supporting the safer use of these technologies in clinical practice.