Paragon Health Institute Proposes Framework to Address AI Medical Device Safety

Paragon Health Institute has published a research paper proposing a new framework to address safety concerns in medical devices that use artificial intelligence, announced in a press release. The paper, titled Generalization Uncertainty in AI-Enabled Medical Devices: A Safer Way Forward, introduces a voluntary method called Digital Similarity Analysis.

Digital Similarity Analysis is designed to help physicians assess whether a patient's data is similar to the data used to train and test an AI device. If the analysis indicates that a patient's data is an outlier, the physician can decide to avoid using the device, request additional validation, or proceed with lower confidence in the device's output.

The framework aims to give physicians better insight into how reliably an AI device will perform for individual patients while allowing manufacturers to maintain the confidentiality of their training data. It also shifts the focus of algorithmic bias evaluation from broad demographic groups to individual characteristics.

The proposal is part of Paragon Health Institute's Health Care AI Initiative, which explores applications of artificial intelligence to improve medical safety, reduce costs, and support innovation in patient care.