Physician in the Loop

Medicare official says parts of AI prior-authorization pilot are not working

Show Notes

Medicare official says parts of AI prior-authorization pilot are not working

The voices on this episode are generated, and so is the script, from the day's post at Physician in the Loop. Every fact in the post links to its source.

Read the post and its sources

Get the free email briefing

A senior Medicare official said Oct. 15 that "there are a lot of things about WISeR that are not working," referring to the agency's pilot of artificial intelligence-supported prior authorization in six states, Nextgov/FCW reported.

Kimberly Brandt, a deputy administrator of the Centers for Medicare & Medicaid Services, made the remark at a conference in Santa Fe, New Mexico, where her slides said the technology requires continuous oversight and that human clinical review is essential. Elsewhere, the Food and Drug Administration cleared an automated insulin pump for adults with insulin-treated type 2 diabetes, and three health systems said they have ended or are ending hospital-at-home programs.

In this episode

What is Physician in the Loop?

How technological innovation is changing medicine for patients and the people who practice on the front lines.

Voice 1: This is Physician in the Loop, the daily briefing on AI in medicine. It's Monday, October 19th, 2026.

Voice 1: A deputy administrator at the Centers for Medicare and Medicaid Services, Kimberly Brandt, said on October 15th that “there are a lot of things about WISeR that are not working.”

Voice 2: That's the AI prior-authorization pilot. What does she actually say is broken?

Voice 1: She made the remark at the ADAPT conference in Santa Fe, New Mexico. Her slides said the technology “requires continuous oversight,” and that human clinical review is essential. She also said Innovation Center staff “are committed to trying to get that right and get it better.”

Voice 2: Who actually has to live with this pilot while CMS works on fixing it?

Voice 1: WISeR started January 1st in Arizona, New Jersey, Ohio, Oklahoma, Texas and Washington. Contractors use AI to review prior authorization requests for things like nerve stimulation, pain-related steroid injections, cervical fusion, knee arthroscopy for osteoarthritis, and skin substitutes, with a clinician reviewing every denial, according to the Congressional Research Service. Providers and beneficiaries cannot opt out.

Voice 2: Has Congress tried to stop it?

Voice 1: A House committee advanced a provision to end the model back in June, but the Senate has rejected efforts to stop it.

Voice 2: Brandt mentioned CMS trying AI somewhere else too, didn't she?

Voice 1: She did. CMS is also piloting AI to flag suspicious Medicaid claims before they're paid.

Voice 2: Hospital-at-home had a rough week of its own, even with Medicare's support still in place. Cone Health in Greensboro said on October 14th it's winding its program down this quarter, citing “evolving reimbursement models, operational complexity and long-term sustainability.” It opened in April 2025, never took more than six patients at a time, and reported no deaths or hospital-acquired infections in its first year.

Voice 1: Is Cone Health the only one stepping back?

Voice 2: No. Health First in Rockledge, Florida is ending its program as part of cutting 214 positions, and Covenant Health in Knoxville, Tennessee already ended its program back in April.

Voice 1: All this while the federal waiver that pays for it is still running?

Voice 2: Right. Congress extended Medicare's Acute Hospital Care at Home waiver through September 30th, 2030. These systems are choosing to step away on their own, not being pushed out by Washington.

Voice 1: The FDA cleared a new option for people managing diabetes at home. It cleared Sequel's twiist automated insulin delivery system for adults with insulin-treated type 2 diabetes, extending a system already used for type 1.

Voice 2: What's the evidence behind it?

Voice 1: In a 13-week study of 307 adults at 30 sites, A1C fell an average of 0.7 point from a baseline of 8.1%, and time in range rose from 57% to 73%, with no severe hypoglycemia, according to Sequel.

Voice 2: Does it ask people to count carbs, like a lot of these systems do?

Voice 1: Barely anyone in the study had done that before. Only 11% of participants had carbohydrate-counting experience, the company said. The algorithm is based on the open-source Tidepool Loop and adjusts basal insulin continuously, and the pump works with Abbott's FreeStyle Libre 3 Plus and Senseonics' Eversense 365 sensors.

Voice 2: A UK digital GLP-1 program posted strong numbers of its own this week. More than 8,000 women in an observational study lost an average of about 16 kilograms over six months, almost all of them on tirzepatide.

Voice 1: How well did people actually stick with it?

Voice 2: About 91% were still in the program at six months, and about 83% had lost at least 10% of their body weight. The authors reported no serious treatment-related adverse events.

Voice 1: Any caveats worth flagging there?

Voice 2: Several. The authors themselves said interpretation is limited by the observational design, the lack of a comparison group, a self-pay population, and just six months of follow-up. The lead author is at University College London, and co-authors are affiliated with the online providers SheMed and eMed Healthcare UK.

Voice 1: Before we go, something to watch.

Voice 2: What's moving on the prior-authorization fight beyond WISeR itself?

Voice 1: State legislatures are moving. Alabama's SB 63, in effect since October 1st, requires a licensed physician or other health care professional to make any prior authorization denial based on medical necessity. Pennsylvania's HB 1925, which would require a clinician's independent review before AI-assisted denials, is awaiting the Senate.

Voice 2: And the payers themselves?

Voice 1: UnitedHealth told investors on October 13th that AI-enabled research cut its appeal cycle times 26%, by its own figures. Centene says it will launch a national prior-authorization gold-card program for 160,000 eligible providers in November.

Voice 2: There's also a deadline coming up outside health care's usual channels.

Voice 1: Comments close October 26th, at 11:59 PM Mountain, on Colorado's automated-decision and chatbot safety rules.

Voice 2: Worth watching whether that reach extends into health care chatbots, given everything CMS just admitted about its own AI tool.

Voice 1: Sources and corrections are at physicianintheloop.org, where you can sign up for the free email briefing. This briefing was read by AI voices.