The Food and Drug Administration has cleared scores of AI tools for use in the clinic, but there remains little publicly available data on how well those tools work on different groups of patients, a STAT investigation has found.
Of 161 AI products cleared by the FDA in recent years, only 73 disclosed the amount of patient data used to validate the performance of their devices in public documents. Only seven reported the racial makeup of their study populations, and just 13 provided a gender breakdown.
Experts say those evidence gaps make it difficult to determine whether AI products will improve outcomes or potentially lead to unintended consequences, such as unnecessary treatments or biased care.
About the Author Reprints
