Month: April 2026

Medicare’s None the WISeR

Posted on by Jason Barry

Washington state just delivered an unfortunate crash course on U.S. health policy after the model aimed at “Wasteful and Inappropriate Service Reduction” led straight to higher costs and fewer treatments for seniors.

Does Medicare need prior authorizations? CMS designed WISeR to find an answer by testing whether bringing AI-driven prior auths (which are already widespread among private payors in Medicare Advantage) to traditional Medicare could cut down on wasteful spending.

  • The six-year pilot kicked off in six states on January 1st (AZ, NJ, OK, OH, TX, WA), targeting a list of 13 “low value” services with a high potential for fraud or waste – most notably orthopedic pain management procedures and skin substitutes.

Washington is already tapping out. Less than five months in, Senator Maria Cantwell (D-Wash.) had enough data to publish her new report on the “clear risks of AI in Medicare.”

  • Drawing on a Washington State Hospital Association survey of 16 hospitals, the report found that procedures previously approved within days are now taking 4 to 8 weeks. 
  • CMS’ own WISeR standards call for responses to providers within 1 day for urgent care and 3 days for routine care, both of which are now clocking in at 15 to 20 days.

You get what you pay for. WISeR compensates third-party administrators for each claim they deny, under the assumption that these denials account for the reduction in wasteful spending.

  • That obviously creates some adverse incentives, which the report eloquently framed up by saying the model “incentivizes WISeR contractors to weaponize AI-driven medical determinations not for the sake of efficiency… but to maximize profitability.”
  • As a result, Washington hospitals have had to add staff and increase hours to manage the surge in prior auths – not a great formula for lowering the cost of care.

The report went straight to the top. At a Senate hearing last week, Senator Cantwell made her case directly to HHS Secretary RFK Jr., who said “that kind of delay is unacceptable.”

  • He went on to say that prior auths are there to prevent the government from being “ripped off” by unethical providers and only applies to 5% of services in Medicare.
  • That might be accurate, but it doesn’t mean they aren’t high-volume services. A separate KFF analysis found that 86% of the 1.1M Medicare beneficiaries that used at least one of the services on WISeR’s list in 2024 received a pain management service.

The Takeaway

Reducing waste in Medicare is a worthy goal, but so far it looks like the best way to make it happen probably isn’t by adding prior auths to the program that many seniors specifically chose to avoid them.

OpenAI Launches ChatGPT for Clinicians

Posted on by Jason Barry

OpenAI is doubling down on providers, and this time around it’s going direct-to-docs with ChatGPT for Clinicians.

ChatGPT-5.4, but for clinicians. OpenAI’s medically-tuned version of ChatGPT uses the same engine as GPT-5.4, with specialized training to optimize it for clinical workflows and administrative tasks.

  • Better yet, it’s free for any verified physician, NP, PA, or pharmacist in the U.S.

ChatGPT for Clinicians includes:

  • Access to GPT-5.4 and OpenAI’s other models (increased limits for clinical tasks).
  • Skills for repeatable workflows (like drafting referral letters and patient instructions).
  • Clinical search based on “millions of peer-reviewed sources” (details were sparse).
  • Optional support for HIPAA compliance (through a BAA for eligible accounts).

ChatGPT for Clinicians doesn’t include:

  • Connectors to the CMS Coverage Database, NPI Registry, or ICD-10. Not impossible.
  • Broader platform advantages like integrated drug information or telehealth.
  • Evidence from big name partners that’s contextualized to the encounter.
  • HIPAA compliance out-of-the-box.

All that said, the performance talks. ChatGPT for Clinicians was launched alongside HealthBench Professional, OpenAI’s new benchmark built from real clinical conversations.

  • It grades models on chat tasks across three use cases (care consults, documentation, and medical research), with physician-authored scenarios and rubrics, as well as scoring designed to reflect real-world performance.
  • As you might expect, OpenAI’s new tool did great on OpenAI’s new benchmark. ChatGPT for Clinicians outperformed Claude Opus 4.7, Gemini 3.1 Pro, and even specialty-matched physicians with unlimited time and web access.

What happened to ChatGPT for Healthcare? OpenAI launched ChatGPT for Healthcare just a few short months ago to help health systems with enterprise-wide deployments, but now it’s taking “the next step” by bringing ChatGPT straight to individual clinicians so that “AGI benefits all of humanity.”

The Takeaway

Whether OpenAI is pursuing an altruistic mission or hedging slow progress in the enterprise arena, millions of clinicians already use ChatGPT to support their care, and now they have a fine-tuned version that won’t break the bank.

PHTI: AI Reality Opposite of Expectations

Posted on by Jason Barry

AI promised less friction and lower administrative costs, but a new report from the Peterson Health Technology Institute suggests that it might actually be delivering the exact opposite.

The report stems from a stakeholder workshop that PHTI held to uncover AI’s impact on two of healthcare’s most hotly debated administrative processes: prior authorization and medical coding.

The main finding highlights an obvious predicament. Speeding up flawed processes doesn’t make them any less expensive. PHTI didn’t pin the blame on either side of the AI arms race.

  • It found that payors are (mostly) using AI responsibly. They’re accelerating PA reviews and auto-approving more clean cases, while simultaneously improving code validation and risk adjustment – although the DOJ would probably disagree.
  • Providers are also using their AI superpowers for good. They’re automating the PA workflows driving burnout and streamlining the coding processes that take clinician time away from patients.

That almost sounds like it should create some efficiency. The problem is that it’s the system that’s broken, and AI doesn’t fix the underlying issues.

  • The report pointed out how AI tools for providers caused an uptick in billing intensity, which payors naturally responded to with across-the-board downcoding and other reimbursement reductions.
  • AI might also reduce the cost for individual orgs to execute or appeal prior auths, but it won’t impact costs for the overall system if nothing gets passed on to patients.
  • PHTI believes this makes reimbursement policy the strongest lever that can realistically be pulled to slash system-level spending.

Follow the incentives. Or in this case, the lack thereof. 

  • On paper, payors and providers should be competing for a finite pool of patients in an arena that rewards better products with smaller price tags. If AI cuts costs, providers would be able to bill less and payors could lower premiums.

Efficiency doesn’t translate to deflation. Payors or providers are rational market actors, and if AI can streamline a process that lets them hold onto more of their revenue, then that’s exactly what they’ll do.

The Takeaway

Bots arguing with bots might be faster than humans arguing with humans, but PHTI doesn’t see that eliminating friction from the overall system if nobody has any incentivize to make it happen. 

The Rise of the Generalist-Specialist

Posted on by Jason Barry

Healthcare’s tidy hierarchy of specialties was formed by cognitive necessity. The corpus of medical knowledge is too massive for a single person to master and the clinical workforce was organized around it, but a new article in Health Affairs says it might be time for a redesign if AI removes that constraint.

AI is scaling specialist-level knowledge. Leading models are coasting through Board exams and polishing their clinical capabilities, which the authors argue will quickly scale specialist-level knowledge to the point where most specialty care can be delivered by PCPs.

  • They coined the term “generalist-specialists” for a new category of doctors that transcends narrow specialty definitions.

Clinical expertise is increasingly democratized. The authors see a future where AI-augmented clinicians can manage the full constellation of patients’ chronic conditions within disease-based domains rather than organ-specific specialties. They give a few examples:

  • Cardiometabolic Diseases – combines cardiology, endocrinology, and nephrology
  • Infectious & Inflammatory – rheumatology, infectious disease, & gastroenterology
  • Primary Care: spans OB/GYN, internal medicine, and pediatrics.

That could have some major benefits. Instead of shuffling a diabetic patient between an endocrinologist, cardiologist, and nephrologist, a generalist-specialist could manage the full cardiometabolic picture.

  • That means fewer handoffs, faster diagnoses, and lower co-pays. It would also unlock a ton of specialty capacity for the patients that need it most.
  • Consolidating care under fewer clinicians would also be a tailwind for value-based care, although it would likely increase utilization in a fee-for-service world by converting deferred, fragmented, or incomplete care into a cohesive billable treatment.

AI isn’t the only barrier to making that happen. Everything from med schools and malpractice standards to credentialing and referral systems would need to be completely overhauled.

  • The generalist-specialist vision also assumes that specialists will be on board with either becoming quasi-PCPs or upskilling to ultra-complex care. Definitely not a given.
  • Patient safety concerns also go without saying, but the AI will probably be pretty decent by the time we have cardio-endocrin-nephrologists putting together the care plans.

The Takeaway

AI could easily bring specialist knowledge to generalist fingertips, but if overworked PCPs are going to start also being OB/GYNs it will take more than a fancy LLM to get there.

Why AI Vendors Struggle to Compete With EHRs

Posted on by Jason Barry

Anyone who has ever tried selling AI into health systems will tell you that it’s tough to compete with EHRs, but a new article in JAMA makes the case that it’s actually gotten too tough – and it might be time for regulators to step in.

Most markets reward the best products. The healthcare industry has a funny way of preventing that from happening, and EHR vendor dominance is a textbook example.

  • EHRs hold advantages across infrastructure, workflow integration, procurement, and pricing that make it difficult for third-party tools to gain a foothold.
  • A 2025 Health Affairs study backed that up by showing that 79% of U.S. hospitals use AI models from their EHR vendor, compared to just 59% that use AI from third-party developers.
  • A Bain report drove the point home. Two-thirds of Epic customers said they’d pick a “good enough” Epic option over a better competing product.

These EHR advantages are a natural feature of the market. That said, it’s up to regulators to decide whether the status quo is serving patients and the overall healthcare system. The JAMA authors argue that it doesn’t, and offer three areas where targeted policy could level the playing field.

Infrastructure – Integrating AI tools into clinical workflows requires real-time data access and the ability to survive EHR upgrades intact, both of which are dramatically easier for EHR vendors – particularly as data fields get added or removed.

  • Potential Policy – Mandate broader API adoption so third parties can access EHR data on equal footing, and use existing EHR certification and interoperability frameworks to do it.

Workflow and Usability – The authors specifically flag EHR vendors’ edge in understanding the trade-offs of allocating limited screen real estate to new AI tools, something that’s harder for third parties to gauge from the outside looking in.

  • Potential Policy – Require EHR vendors to offer more robust developer sandboxes – similar to Apple’s iOS developer environment – so third parties can build and test without operating at a structural disadvantage.

Procurement and Pricing – Long-standing health system relationships give EHR vendors a streamlined path through procurement, as well as the leverage to “use pricing structures that incentivize adoption.”

  • Potential Policy – Although this is the hardest area for a policy fix, the authors suggest that improving transparency around AI performance could at least help health systems make more informed decisions regardless of where a tool comes from.

The Takeaway

EHRs are in a powerful position, and companies in powerful positions have a long track record of making life harder for their competition. Healthcare is too important of an industry to not have the best products rise to the top, and this article offers some sound strategies to make sure that stays possible.

Rock Health Q1: Capital Continues Concentrating

Posted on by Jason Barry

Spring is finally here, and Rock Health’s Q1 funding recap shows that the investing landscape is definitely looking greener than last year.

Digital health startups raised $4B on the dot. That’s a whole billion higher than Q1 2025, although the gains were far from evenly distributed.

Here’s Q1 2026 by the numbers:

  • Digital health funding totaled $4B across 110 rounds (vs. $3B and 122 rounds last year).
  • Average round size climbed to $36.7M (highest since Q4 2021).
  • Rock Health counted 12 mega-rounds over $100M.

That last bullet defined the quarter. A dozen companies accounted for 59% of all capital deployed in Q1, one of the highest concentrations Rock Health has ever seen.

  • Round sizes have consistently increased every quarter since 2024, and there haven’t been this many nine-figure checks in a quarter since the pandemic peak in 2022. 
  • Whoop landed $575M at a $10B valuation, Verily raised $300M as it steps out from under Alphabet’s umbrella, and OpenEvidence’s fundraising blitz added another $250M.

The check sizes only tell half the story. One of the reasons why startups are raising bigger late-stage rounds is because they’re waiting longer to go public. 

  • Hinge and Omada broke the ice, but all it took was a little “geopolitical uncertainty” to spook investors and close the IPO window right behind them.
  • If the rest of 2026 pans out like the first quarter, we’d see close to 50 mega-rounds, almost double last year’s count.

AI is now the operating environment. The tech has become so ubiquitous that Rock Health said it will no longer be using “AI-enabled startups” as a distinct category in its funding reports.

  • The broader market remains bullish on the value of AI, but if everyone has it then it stops being a differentiator.
  • The AI startups successfully raising are the ones moving earliest into complex use cases, like Doctronic’s prescribing pilot in Utah or Qualified Health’s governance platform for health systems.

The Takeaway

Q1 mostly brought more of the same. Investors are active but selective, and the chasm between the Davids and Goliaths isn’t getting any smaller. AI is helping startups move faster than ever, but the rest of the year should help clarify whose momentum is actually durable.

Inside the World’s First One-Man AI Unicorn

Posted on by Jason Barry

We officially have our first one-man, billion-dollar AI startup, and it’s in healthcare… for better or worse.

It was hard to be online last week without stumbling across the New York Times profile on “How A.I. Helped One Man (and His Brother) Build a $1.8 Billion Company.”

The article tells a shimmering tale of AI entrepreneurship. The 41-year old founder took just two months, $20k, and a dozen AI tools to get his startup Medvi off the ground.

  • Medvi plays at the intersection of two trends that seem almost engineered to mint new billionaires: agentic AI and GLP-1s.
  • The online telehealth provider offers GLP-1s for weight loss, with an army of AI agents handling everything from website copy and design to ad images and customer service.
  • It had 300 customers in its first month, generated $401M in its first full year (2025), and revenue is now on track to hit $1.8B after the founder doubled the headcount by hiring his brother.

Medvi is basically dropshipping healthcare. They’re in the business of acquiring customers, not delivering care. The entire clinical infrastructure runs through turnkey partners.

  • CareValidate manages the physician licensing and prescriptions.
  • OpenLoop handles the pharmacy fulfillment and shipping.

Here’s what NYT forgot to mention. The article notes that Medvi’s founder was “nervous to talk publicly” about the company and hasn’t exactly been waving his momentum around.

Some solid sleuthing from digital health’s finest offers a few hints as to why that might be:

  • Medvi (or partners with a conveniently long leash) spun up 800+ fake doctor Facebook accounts to aggressively advertise their meds. Not great.
  • They’re named in a lawsuit alleging a nationwide scheme to manufacture and promote a fraudulent, unapproved oral tirzepatide pill. Also not great, allegedly.
  • Their onboarding accepted a user with a February 31st birthday, then told them they had a 94% chance of hitting their goal weight of 200lbs starting from 7’11” and 350lbs. Yikes.

Sam Altman called it. The OpenAI CEO was spot on with his prediction that AI would quickly let solo founders generate billions by eliminating the inefficiency seen at larger orgs. Unfortunately in healthcare, much of that “inefficiency” is in place to protect patients.

The Takeaway

Is it impressive that one founder and a little grayhat marketing can now do billions in revenue? Yes. Should we be dropshipping healthcare through predatory ad funnels? Probably not. 

Scribes Show Modest Impact at Major Academics

Posted on by Jason Barry

Ambient scribes are back in the spotlight after a new study in JAMA confirmed that they move the needle on productivity metrics, but the jury’s still out on whether that’s the best yardstick for success.

This was a big one. The study examined the impact of AI scribe use on over 1,800 clinicians at five major academic medical centers from 2023 to 2025.

  • The academics: MGB, YNHH, UCSD, UCSF, UC Davis 
  • The scribes: Abridge, Ambience, Microsoft DAX Copilot

Here’s what they found. Clinicians who used AI scribes:

  • Saved 16 minutes of documentation time per eight hours of patient care 
  • Saved 13 minutes of EHR time 
  • Could see one additional patient every two weeks
  • Saw no significant impact on EHR timeoutside of working hours

Usage patterns helped color in the story. While 1,800 AI scribe adopters is one of the largest samples out there, the 6,770 control clinicians were also offered scribes and opted not to use them.

  • The biggest gains went to the biggest users. Clinicians who used the AI scribe for over 50% of visits experienced twice the reduction in EHR time and 3x the reduction in documentation time, yet only 32% of adopters fell into this bucket.

What’s counted? What matters? This isn’t the first study we’ve covered that scores AI scribes based on metrics that researchers can easily measure (EHR time, visits), which isn’t necessarily the same as the metrics that matter most to patients or clinicians.

  • Although this study solidifies that scribes can cut documentation time, the question now is if that time gets reinvested in ways that improve care and outcomes for patients.
  • The results also confirm that the mechanism of action for scribes reducing burnout isn’t through time savings, but it’s still unclear whether it’s from having a couple more moments to take a deep breath throughout the day or from reallocating the extra minutes to things that feel valuable.

The Takeaway

This study offers the most definitive real-world data yet that AI scribes have a modest impact on productivity metrics, but it also confirms that cleaner notes aren’t the only key to improving healthcare experiences.

Get the top digital health stories right in your inbox