CMS tech rules and AI benefits for primary care, patients and ‘touchless’ prior auths

The concept of using artificial intelligence (AI) to streamline the prior authorization process sounds like a great idea.

What will that look like in an actual medical practice? And how will patients respond?

The U.S. Centers for Medicare & Medicaid Services has set Jan. 1, 2027, as the date to to comply with CMS-0057-F, the CMS Interoperability and Prior Authorization Final Rule, which requires health plans to run four standardized electronic systems, including one built specifically for prior authorization requests.

Cohere Health joined the new CMS Electronic Prior Authorization Acceleration Initiative, a coalition of health plans, providers and technology vendors working to meet that deadline. Company Chief Medical Officer Brian Covino, M.D., spoke with Medical Economics about which specialties stand to benefit most and what a fully electronic, AI-assisted practice might look like. Patient reaction has been good, he said, though healthy skepticism is warranted. That skepticism emerged this year in results of an American Medical Association survey, but it will be allayed when physicians and patients know that doctors, not just technologists, have built the electronic tools used in health care.

This transcript has been edited for length and clarity.

Which medical specialties do you see benefiting most from more electronic communication and AI in prior authorization?

Brian Covino, M.D.: For health plans, a lot of the bulk of utilization management and prior authorization is in areas such as musculoskeletal medicine, myself being an orthopedic surgeon, cardiology, imaging and oncology. Those are the areas that often have significant utilization management efforts. We've partnered with some specialty medical societies, such as the American Academy of Orthopaedic Surgeons and the American College of Cardiology, to use the appropriate-use guidelines those societies have created and embed them into our decision-making. We can codify their policies and digitize them, so that when we're extracting information, if a request meets those guidelines, it can get approved. I think AI will be very helpful in those areas because they're high-volume. Imaging, for example, is a very high-volume area for prior authorization. Often for procedures that aren't much of a patient-safety issue, just high-volume, low-cost care where AI has really helped speed up patients getting an MRI, an image or a CT scan more quickly.

For primary care physicians specifically, I understand they're seeing high volumes of complex patients with lots of problems and don't have as much time as they need to care for them. Any way we can make this process easier for them matters, because they do submit a significant portion of prior authorization requests. Standardization would help them and their staff, too. My message to primary care physicians is: We appreciate what you do. Help is here, and help is on the way, to try to make your lives easier for the tremendous number of patients you care for on a daily basis.

What will the technology needs of a medical practice look like going forward — a proprietary program, a subscription service? How do you envision that?

Brian Covino, M.D.: I think the future state of this is what we'd call touchless authorization. Most primary care physicians' offices now have electronic health records (EHRs), so ideally, while a patient is still sitting with the physician and the physician says, ‘We ought to get an ultrasound or a CT scan of this,’ they could enter that order into their system, and that would immediately start the prior authorization process through the EHR. We could connect to that, and once we're notified, we could start pulling the information out of the EHR. Right now, physicians and their staff have to push the information to us. If we could be interoperable and pull it from the EHR or from the health plan's claims on that patient, to make sure they haven't already had that done somewhere else, we could get the approval while the patient is still sitting in the office. Think about a patient who drives an hour or two to see a doctor and is told, ‘You need a test done, come back in two weeks once I find out if it's approved.' Compare that to, ‘It just got approved, let's send you next door and have this test done right now.' I think that's the future of it, where we can make this touchless, and continue to increase efficiency and reduce delays.

Surveys have found patient skepticism about AI in health care and administrative decisions. Have you encountered that? What's the patient reaction been?

Brian Covino, M.D.: We don't talk directly to patients, but we talk to the practices treating them, and it's been pretty uniformly well received. I think the reason is that patients are getting approvals very quickly for appropriate care, so some of that initial skepticism, which is understandable, creates less concern once they see their doctor scheduling their procedure or test faster, and getting answers to their questions instead of sitting at home wondering what's wrong.

Some of that skepticism is somewhat healthy, though, and it goes back to educating physicians, clinicians and patients about the appropriate use of this technology. Think about it: A doctor, or a doctor like your doctor, is sitting there developing this alongside the technologists. I love our tech colleagues in Silicon Valley and places like that, but this isn't just a bunch of engineers building this on their own. Knowing that it's your physician, or somebody just like your physician, developing this should hopefully be a comfort that it'll be accurate and beneficial.

Physicians have said EHRs were imposed on them by insurers without much input. How important is physician involvement in building AI tools like this?

Brian Covino, M.D.: I agree, and I'll tell you that from the very beginning at Cohere, even when this was a concept developed within a health plan before it became a company, there was physician involvement from the start. To get buy-in from the medical community and from patients, they have to understand that clinicians have been involved in the process. Those of us who had been in practice knew the problems with the process; we could tell you very clearly what the issues were, so we knew where to go and what to try to solve. You can't overstate how important that is in the development, both in terms of the technical accuracy and in knowing that doctors are solving the problems for themselves here. Not everybody has the time to do that, but fortunately I did, and others have as well, to try to solve some of these problems for our colleagues and for the patients we care for.

What didn't I ask about that you'd like to emphasize, especially about the CMS rules and the shift toward more AI in prior authorization?

Brian Covino, M.D.: What I would say is that there are people out here in the health care world trying to make this better. It's an onerous process, and it's probably not going away, because unfortunately we still have a small percentage of care delivered in this country that's wasteful, inappropriate and part of why health care is so expensive. There are people working on this. Health plans really want to change it, too, and CMS and the Center for Medicare and Medicaid Innovation really want to solve this for Medicare and Medicaid beneficiaries as well. There are good people, both in private industry and in government, working to solve this problem. It's not easy, and it takes time, but there are a lot of good people trying to make this better, because we all want to make sure our health care industry stays the best, and hopefully becomes less expensive than it is currently.