Why medical AI and value-based care may need each other

According to recent estimates, the Food and Drug Administration has approved 950 medical devices using artificial intelligence and machine learning. However, far fewer have had a significant impact on patient care in the United States.

For medical AI developers like James Zou, this is a problem. A Stanford professor is investing in effective medical machine learning – including some of his own FDA-approved algorithms – for healthcare systems.

In two recent articles, Zou examined the clinical use of medical AI devices, finding that a small minority of FDA-approved devices are reimbursed at any scale, and delving into various models for how health systems could pay for medical AI.