By KIM BELLARD
The Blue Cross Blue Defend Affiliation is crying wolf about AI. Its new analysis claims that “hospital methods are more and more billing affected person hospital stays as extra medically advanced,” and that these upcodings price BCBS plans virtually $1b from 2023 to 2025. That’s simply claims from hospital methods, simply to BCBS plans, so the entire quantity may very well be a lot greater.
Nicely, you would possibly say, a billion right here, a billion there – that’s simply healthcare as of late. All the things goes up. However BCBSA is much less sanguine; it thinks the coding is a part of a method, and it thinks AI is driving that technique.
“If sufferers are actually sicker, we’d anticipate to see extra therapy,” mentioned Luke Chalker, BCBSA’s senior vp of product and information science. “For instance, we’re seeing considerably extra anemia diagnoses at these hospitals with no corresponding improve in transfusions. The disconnect between diagnoses and therapy means that AI is figuring out extra billable circumstances, not sicker sufferers.”
The evaluation indicated that almost all (70%) of the elevated prices got here from secondary diagnoses that shifted fee into greater reimbursement classes, so the speculation is believable.
The American Hospital Affiliation, in fact, is having none of it. An August 2026 ”fact sheet” on AI and coding depth known as claims of AI driving coding depth “unsubstantiated,” citing quite a few elements that drive coding depth. It blamed insurers for having AI-supported “downcoding applications,” the aim of that are “to arbitrarily and inappropriately cut back supplier reimbursement for medically needed care that has already been delivered.”
AHA additionally snarkily however precisely famous what number of Medicare Benefit plans have been caught upcoding members to spice up their Medicare funds, so it is dependent upon the place one is sitting.
It’s additionally value noting that this 12 months the Trump Administration rolled out a Medicare program — Wasteful and Inappropriate Service Discount (WISeR) – to make use of AI to assist determine “fraud, waste and abuse” – a.ok.a., deny claims. It has not, to say the least, gone nicely.
Each side are utilizing AI, to one another’s detriment. Which is to say, our detriment.
The BCBSA press launch notes that greater than 60% of hospital methods are utilizing AI applied sciences, and, certainly, the 2026 Oliver Wyman Healthcare RCM Survey discovered that 63% of well being methods are utilizing AI of their income cycle administration. However it’s not fairly the smoking gun BCBSA would have us imagine. Oliver Wyman discovered that “roughly 20% to 40% of organizations we surveyed report broad or enterprise-wide use of AI-enabled instruments,” however this included ambient listening (33% in broad use or enterprise-wide commonplace), scientific choice assist (38%), scientific doc integrity (41%), digital prior authorization (41%), and coding/billing automation (43%).
Alternatively, Oliver Wyman additionally discovered 70 to 90% of choice makers anticipated to extend AI-related RCM spending over the subsequent three years, and it’s a must to marvel why that quantity isn’t 100%.
To BCBSA’s level, the report did discover that well being methods “are seeing significant positive factors with focused makes use of of AI enhancing how scientific complexity is captured, with some research displaying accuracy hit 90% or greater in particular scientific domains,” with the end result that “reimbursement ranges are rising, contributing to measurable shifts in cost-of-care traits.”
Accordingly, the report urged “payers have to speed up investments in analytics, fee integrity capabilities, auditing and validation.” Which can drive higher investments from well being methods, and so forth.
And, in case you have been questioning, it isn’t simply well being methods and payors utilizing AI. The American Medical Affiliation’s “2026 Physician Survey on Augmented Intelligence” discovered that greater than 80% of physicians are utilizing AI professionally, double what was reported in 2023. Like well being methods, they’re utilizing it for many causes, however “diagnostic accuracy” is listed as one of many best anticipated benefits. Whether or not “accuracy” means extra precision or upcoding is dependent upon which aspect of the ledger one is on.
Yep, we’re in an AI struggle about healthcare claims, though Mr. Chalker says: “It’s not a struggle. It’s a totally one-sided blood bathtub.”
In a New York Times article on the BCBSA report, Reed Abelson and Teddy Rosenbluth quote Dr. Shiv Rao, CEO and founding father of healthcare AI agency Abridge: “That’s the hazard, the place it’s bots combating bots, brokers combating brokers, a horrible dystopic future no person desires to stay in.” In addition they cited a current panel with leaders from Cigna and Emory Well being, throughout which Dr. Joon S. Lee, the chief govt of Emory Healthcare, mentioned: “My nightmare is that now we have bots on Emory Healthcare’s aspect, bots on Cigna’s aspect that hold speaking to one another. An entire information heart may very well be occupied simply doing that.”
It ought to be everybody’s nightmare. Bots on bots doesn’t make sufferers higher. Bots on bots doesn’t make healthcare extra inexpensive. Bots on bots simply exacerbate the issues we have already got in healthcare.
Ms. Abelson and Ms. Rosenbluth level to a Well being Affairs article earlier this 12 months by Dr. David Brailer, a former Obama official who helped oversee the adaption of digital well being information (EHRs). Its title — Why AI Will Accelerate Health Care Inflation – is daunting. EHRs, he factors out, led to greater hospital funds, and we must always anticipate the identical with AI. “The extraction financial system of American well being care doesn’t resist helpful expertise,” he mentioned. “It recruits it. AI is being recruited the identical means, and AI’s scale and velocity is totally different from something digital information achieved.”
In different phrases, if we’re not very cautious and really considerate, AI will make all of the issues we hate about our healthcare system a lot worse, a lot sooner. And we don’t have a lot observe file in being very cautious or very considerate.
At Well being Datapalooza, Amy Gleason, the Administrator of the U.S. DOGE Service, had an excellent quote: “We’ve got normalized the absurd in healthcare.” I couldn’t agree extra. If AI serves to additional normalize what everybody agrees are the absurdities of our healthcare system, then perhaps it’s time for AI to do away with us.
I’ve lengthy argued that every one of our issues in healthcare stem from the truth that we don’t actually know what “high quality” is for well being care. Due to that, we additionally don’t actually know who’s offering it, or who is just not, both deliberately or unintentionally. Focusing AI on that might profit each those that present care and those that pay for it, to not point out those that obtain it.
Sure, AI ought to assist us make our healthcare system extra environment friendly, decreasing the ridiculous administrative duties now we have allowed to build up, but when that simply ends in “bots on bots” combating over prognosis codes, we’ll have wasted our greatest probability to truly repair our healthcare system.
Kim is a former emarketing exec at a significant Blues plan, editor of the late & lamented Tincture.io, and now common THCB contributor
