House Democrats are pressuring the Centers for Medicare & Medicaid Services (CMS) to better evaluate how artificial intelligence is utilized by Medicare Advantage (MA) plans, hoping to prevent eye-catching rates of denial.
In a letter (PDF) released Friday and endorsed by the Center for Medicare Advocacy, Business Leaders for Health Care Transformation and other groups, 30 House members led by Reps. Judy Chu, D-California, and Jerrold Nadler, D-New York, said they are concerned about CMS’ prior authorization requirements in the final 2024 Part C and D rule.
They said MA plans’ use of AI software, managed by firms like naviHealth, myNexus and CareCentrix, has led to restrictive decisions and repeated denials, resulting in a lower quality of care for patients.
“Absent prohibiting the use of AI/algorithmic tools outright, 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 letter reads.
MA plans have said these tools guide care decisions but do not dictate denials. And while CMS has not banned algorithmic-driven tools, the agency has publicly stated that decisions must be made based on medical necessity determinations and not AI software.
“The use of AI by Medicare Advantage insurers to deny needed care to seniors and people with disabilities represents the most recent and dangerous step by greedy companies focused on profit instead of patients,” said Lisa Gilbert, executive vice president at Public Citizen, in a statement. “Now is the time for CMS to crack down on companies that are using AI and other mechanisms to deny care that would be covered if the enrollee were covered by Traditional Medicare. Understanding how Medicare Advantage insurers are using AI to deny needed care and holding bad actors accountable are crucial steps to protecting seniors and the Medicare program,”
The House Democrats provided a list of measures they hope CMS will take which would require MA plans to give reasons (including descriptions of a patient’s condition and the denial timeline) for why a patient is denied service, assess the frequency of denials, determine the extent to which AI played during the denial process, whether algorithms are self-correcting (in that if a denial is reversed on appeal, that is then factored into the algorithm for next time) and whether plans are inappropriately using race or other factors in algorithms.
The group said it also wants AI rulings to be compared against traditional coverage decisions and wants assurances from MA plans and contractors that coverage is not more restrictive than traditional Medicare.
“Medicare Advantage plans are entrusted with providing medically necessary care to their enrollees,” the members said. “While CMS has recently made considerable strides in ensuring that this happens, more work is needed with respect to reining in inappropriate use of prior authorization by MA plans, particularly when using AI/algorithmic software.”
Traditional Medicare does not require prior authorization, but, under MA, prior authorization is rampant. In a 2022 Office of Inspector General report (PDF) cited by House Democrats, it found that 87% of claims that were denied did not meet Medicare coverage rules, meaning the services would’ve been covered under original Medicare.