Senate and House Democrats introduced resolutions Wednesday to overturn a Medicare pilot that uses artificial intelligence to approve or deny care. The technology experiment has led to allegations of delayed care for seniors in at least one of the six states it is currently being implemented.
The newest effort to force a vote to stop the experiment follows a Government Accountability Office ruling from last week. The GAO determined that the Wasteful and Inappropriate Service Reduction program (WISeR) is subject to congressional approval and should have been submitted to Congress before going into effect in January.
“Americans are sick and tired of abusive prior authorization tactics putting needed health care out of reach,” Senate Finance Committee ranking member Ron Wyden (D-Ore.) said in a press release. “The last thing seniors need is even more AI denying the care they need.”
Wyden sponsored the resolution to invoke the Congressional Review Act along with Sens. Patty Murray (D-Wash.), Maria Cantwell (D-Wash.), Richard Blumenthal (D-Conn.), and Kirsten Gillibrand (D-N.Y.). The coalition of 20 Senate Democrats behind the resolution shows a broadening resistance to the program beyond the six states it has been implemented in: Arizona, New Jersey, Oklahoma, Ohio, Texas, and Washington.
Reps. Suzan DelBene (D-Wash.) and Greg Landsman (D-Ohio) simultaneously introduced a similar resolution in the House. The resolution comes after DelBene last year introduced a bill to stop the Medicare experiment, which was accompanied by a similar bill in the Senate.
“WISeR is a dangerous program that is denying care to Medicare patients so companies can profit,” said DelBene in a press release. “This program implements the same flawed prior authorization scheme from Medicare Advantage into traditional Medicare. If scaled up, it would be a back door to privatizing Medicare. It is causing needless delays, worsening conditions, and costing us more in the long run. Congress needs to step up and put an end to WISeR.”
In April, Cantwell reported that seniors in Washington are facing delays in care due to the program. In a congressional hearing at the time, she said that the Centers for Medicare and Medicaid Services is using AI as a “denial device.”
The WISeR program targets a set number of supplies and procedures that are largely associated with fraud, according to CMS, and inpatient-only services, emergency services, and services that would pose a substantial risk to patients if delayed, are excluded. Providers need to submit these supplies and procedures to a set of WISeR contractors that use technology such as AI to approve or deny the care. The contractors are paid based on an undisclosed formula that includes how many procedures they have denied, leading outside experts to warn that the program incentivizes denying care.
DelBene raised concerns about this payment scheme to CMS administrator Mehmet Oz, who said in a recent letter that the agency will be auditing participants to ensure that their decisions are consistent with existing Medicare coverage criteria, and that vendors with a high inaccuracy rate will be terminated.
Reached for comment after the GAO ruling, a CMS spokesperson told STAT that WISeR “remains an active Innovation Center model, and CMS will continue to review the opinion and assess any appropriate next steps consistent with applicable law and administrative processes.”
Votes invoked under the Congressional Review Act have sparingly been used to overturn previously approved rules from federal agencies. Though more than 250 rules have been challenged this way, only a handful have been successfully overturned, most notably 16 rules from the end of the Obama administration that were stopped at the beginning of the first Trump administration.
The resolution needs 30 signatures from senators to bring it out of committee and to the Senate floor for a simple majority vote. It’s unclear whether the movement will garner bipartisan support, but if the resolution passes both chambers, the WISeR pilot will immediately become ineffective.
