Scott Gottlieb Calls for FDA to Revert to Older Policies on Artificial Intelligence
Dive Brief:
- Scott Gottlieb, the former Food and Drug Administration (FDA) commissioner under the Trump administration, recently published an article in JAMA calling for the agency to return to previous policies on artificial intelligence (AI).
- Gottlieb, who led the FDA from 2017 to 2019, expressed concerns about the uncertainties introduced by recent changes in AI regulation policies.
- He suggested reverting to an older interpretation of the 21st Century Cures Act to exempt more types of clinical decision support software from the FDA’s premarket review process.
Dive Insight:
The FDA’s approach to regulating AI in medical devices has come under scrutiny with the transition to the second Trump administration. With no permanent FDA commissioner appointed yet, questions loom over how AI will be regulated in the medical field.
President Donald Trump’s executive order emphasized the need to remove barriers hindering American AI innovation. Experts anticipate potential policy reversals during the Biden administration, including the controversial clinical decision support final guidance mentioned in Gottlieb’s article.
The FDA issued the final guidance in 2022, expanding its oversight to include risk-scoring tools like sepsis or stroke prediction software. Previously, these tools were exempt from premarket review if they did not solely dictate diagnostic or treatment decisions.
Concerns such as automation bias and the integration of data from various sources were addressed in the guidance. The FDA’s decision to regulate a sepsis risk-scoring tool developed by Epic further emphasized the need for stringent oversight.
Gottlieb highlighted the unintended consequences of the final guidance, noting that EMR developers restricted software features to avoid regulatory scrutiny. He proposed a return to older policies that exempt AI tools designed to enhance clinical information without making independent diagnoses or treatment decisions from premarket review.
