83 days on a 72-hour clock: what Medicare’s AI pilot exposes
New records from Medicare’s AI prior-auth pilot show long delays. The more useful question for surgical groups: Do you know your own authorization numbers?
KAIRIX · Mon Nov 09 2026 00:00:00 GMT+0000 (Coordinated Universal Time)
Last week, reporting on records obtained by the Electronic Frontier Foundation gave the clearest look yet at Medicare's WISeR model, the pilot that uses AI vendors to review prior authorization requests for traditional Medicare in six states. Tech Times reported that four of the six vendors drew CMS compliance actions in the program's first months, and that one request went unanswered for 83 days against a 72-hour standard. MedCity News reported that one vendor rejected 53% of more than 6,000 requests over three months. It also noted that the figure alone doesn't show the denials were wrong.
A separate JAMA research letter, covered by TechTarget, found that volumes for targeted services fell about 20% in the pilot states compared with other states after the program began. Epidural steroid injections fell about 15% from January to June. The authors were careful to say they can't yet tell how much of that drop was wasteful care and how much was care patients needed.
I'm not a clinician, and I'm not going to argue the medicine. What I can speak to is the operations side, because I spend most weeks talking with administrators, COOs and practice CEOs at independent surgical groups. This story isn't really about AI. It's about time, and whether a practice can see where its time goes.
What this changes for a surgical group
For a spine, orthopedic or neurosurgical group, an authorization isn't a single event. It's a request, a wait, sometimes a peer-to-peer, sometimes a resubmission, and then a clock running toward expiration. If the case slips because imaging was late, the surgeon's block moved or the patient needed clearance, the authorization can lapse before the surgery date. Then the work starts over. The pilot covers a defined list of services, including injections, nerve stimulators and knee arthroscopy, but payer rules change constantly. Every new requirement lands on the same front-office staff.
The numbers most practices can't pull
So here's the operational question I'd put to any practice leader reading this: what share of your authorizations expire before the case? And right behind it: how long does an authorization take from order to approval, by payer? How many touches does each one take? How many cases were rescheduled because the authorization wasn't in place?
In my experience, most groups can't answer those questions from a report. The data is spread across the EHR, payer portals, a fax queue and a shared spreadsheet someone built years ago. The staff doing the work know it's hard, but they usually can't tell you how hard in numbers, because nobody has asked them to measure it. Not knowing is the finding.
What we learned measuring referrals
We've seen this pattern before with referrals. When we measured the referral process at a 40-surgeon neurosurgical group, the baseline was about 63 manual touches per referral and about 53 days from referral to a scheduled appointment. We got those numbers by tracing the work step by step, from the fax arriving to the appointment on the calendar.
Prior authorization is the next benchmark we're building, and we're starting with the expired-before-the-case rate. That number tells you whether the problem is the payer, the scheduling handoff or the internal queue. Each of those needs a different fix. Treating them as one problem called "prior auth" is how practices end up hiring another coordinator and getting the same result.
What's your number?
The WISeR story will keep moving. The records indicate CMS still plans to expand the model, including a gold-card exemption for highly compliant providers, and lawmakers continue to push back. Whatever happens there, a practice that knows its own authorization numbers will be in a better position to respond than one that's estimating.
If you run operations at a surgical specialty group and you know your expired-auth rate, I'd like to hear it. If you don't, that's useful too. Reply with your number, or with "we don't know," and I'll share what we're seeing as the benchmark comes together.
Sources
Medicare's AI Prior Authorization Pilot Had 4 of 6 Vendors Flagged in First Weeks, Records Show — Tech Times, Nov. 3, 2026
7 Potential Problems with Medicare's AI Prior Authorization Experiment — MedCity News, Nov. 2, 2026
CMS prior auth pilot shrinks service use amid public backlash — TechTarget, Nov. 3, 2026 A separate JAMA research letter, covered by TechTarget, found that volumes for targeted services fell about 20% in the pilot states compared with other states after the program began. Epidural steroid injections fell about 15% from January to June. The authors were careful to say they can't yet tell how much of that drop was wasteful care and how much was care patients needed.
I'm not a clinician, and I'm not going to argue the medicine. What I can speak to is the operations side, because I spend most weeks talking with administrators, COOs and practice CEOs at independent surgical groups. This story isn't really about AI. It's about time, and whether a practice can see where its time goes.
What this changes for a surgical group
For a spine, orthopedic or neurosurgical group, an authorization isn't a single event. It's a request, a wait, sometimes a peer-to-peer, sometimes a resubmission, and then a clock running toward expiration. If the case slips because imaging was late, the surgeon's block moved or the patient needed clearance, the authorization can lapse before the surgery date. Then the work starts over. The pilot covers a defined list of services, including injections, nerve stimulators and knee arthroscopy, but payer rules change constantly. Every new requirement lands on the same front-office staff.
The numbers most practices can't pull
So here's the operational question I'd put to any practice leader reading this: what share of your authorizations expire before the case? And right behind it: how long does an authorization take from order to approval, by payer? How many touches does each one take? How many cases were rescheduled because the authorization wasn't in place?
In my experience, most groups can't answer those questions from a report. The data is spread across the EHR, payer portals, a fax queue and a shared spreadsheet someone built years ago. The staff doing the work know it's hard, but they usually can't tell you how hard in numbers, because nobody has asked them to measure it. Not knowing is the finding.
What we learned measuring referrals
We've seen this pattern before with referrals. When we measured the referral process at a 40-surgeon neurosurgical group, the baseline was about 63 manual touches per referral and about 53 days from referral to a scheduled appointment. We got those numbers by tracing the work step by step, from the fax arriving to the appointment on the calendar.
Prior authorization is the next benchmark we're building, and we're starting with the expired-before-the-case rate. That number tells you whether the problem is the payer, the scheduling handoff or the internal queue. Each of those needs a different fix. Treating them as one problem called "prior auth" is how practices end up hiring another coordinator and getting the same result.
What's your number?
The WISeR story will keep moving. The records indicate CMS still plans to expand the model, including a gold-card exemption for highly compliant providers, and lawmakers continue to push back. Whatever happens there, a practice that knows its own authorization numbers will be in a better position to respond than one that's estimating.
If you run operations at a surgical specialty group and you know your expired-auth rate, I'd like to hear it. If you don't, that's useful too. Reply with your number, or with "we don't know," and I'll share what we're seeing as the benchmark comes together.
Sources
MedCity News, "7 Potential Problems with Medicare's AI Prior Authorization Experiment," Nov. 2, 2026
TechTarget, "CMS prior auth pilot shrinks service use amid public backlash," Nov. 3, 2026