Although it has yet to result in actual legislation, congressional leaders continue to raise concerns about artificial intelligence in healthcare and its impact on denying coverage to seniors.
Thirty House Democrats released a letter Friday urging the Centers for Medicare & Medicaid Services to take several steps to make sure AI is not unfairly denying coverage.
“Absent prohibiting the use of AI/algorithmic tools outright,” the letter reads, “it is unclear how CMS is monitoring and evaluating MA plans’ use of such tools in order ensure that plans comply with Medicare’s rules and do not inappropriately create barriers to care.”
The latest concern is that AI’s contribution to coverage decisions, and coverage denials, is framed as being based on health data when, in fact, the AI is just reinforcing cost considerations. For instance, if a decision is decreed that an older adult with a hip fracture should spend “X” number of days at a skilled nursing facility.
In addition to possibly dictating care coverage (or lack thereof), other concerns about AI related to long-term care include whether AI tools are embedded in systems without regard for transparency, or if AI compromises residents’ privacy by releasing medical histories.
The congressional recommendations for CMS:
- Requiring Medicare Advantage plans to include specific reasons for denial of coverage.
- Determine whether AI is making accurate medical diagnoses — for instance, how long someone should stay in a skilled nursing facility for health reasons. Or is it simply mimicking MA’s prior coverage determinations, which may be more related to limiting costs.
- Assess and report on how AI is contributing to the rate of coverage denials.
Senators from both major political parties also released statements last month regarding concern for how AI is used to determine coverage. Sens. Bill Cassidy (R-LA) and Chuck Schumer (D-NY) indicated support for future legislation that would restrict AI’s ability to contribute to coverage decisions.
More broadly, the White House issued an executive order last week on AI’s use in all fields, including healthcare, although that order called for more evaluation and the formation of guidelines for AI rather than outright restrictions.
