Today’s newsletter is brought to you by: All About Vision

When Humana got 73,000 members with diabetes in for overdue eye exams, nearly one in four exams found undetected eye disease. Eye care rarely comes up in healthcare strategy conversations, but we should keep a closer eye on it — last year, venture funding for ophthalmology more than doubled.

In this brief, sponsored by All About Vision, HTN pairs new Harris Poll data with our analysis of CMS plan files to show where vision care fits for payers, consumers, and startups.

If you're interested in partnering with Health Tech Nerds, let us know!

Programming Note: I’ll be taking next Sunday off for Thanksgiving, but hopefully this 5,000-word edition will leave you with plenty to ponder on the Medicare, primary care, and consumer fronts to tide you over for a few weeks.

An ask of you: please take ~7 minutes to fill out our customer feedback survey! You’ll see the link below, and it’ll help us immensely as we plan ahead for 2027.

Happy Thanksgiving to all my U.S. based friends. I’ll see you again Sunday, 12/6!

Medicare Advantage, Part I

1. A CRUSHing week for Medicare Advantage

The bad news for Medicare Advantage started Monday, when CMS released a new proposed rule, Comprehensive Regulations to Uncover Suspicious Healthcare (CRUSH). Dr. Oz took to social media to release a video explaining how this proposed rule will help CMS crush “fraudsters” and remove ~$90 billion in fraud, waste, and abuse from healthcare over the next decade.

  • Who are these fraudsters? As Bob Herman noted on X, Oz apparently was referring to MA plans as fraudsters, since CRUSH's $90 billion in savings over the next decade would come entirely from Medicare Advantage plan clawbacks via risk adjustment data validation (RADV) audits. Humana has an ongoing lawsuit with CMS over RADV, which may indicate we should expect litigation ahead here as well.

  • The M4A double-whammy. The bad news for Medicare Advantage was compounded this week as Medicare for All conversations continue to pick up steam, led by Bernie Sanders laying out his agenda on Wednesday. Between the combination of CRUSH and M4A conversations, the Medicare Advantage sector did not have a good week on Wall Street, with Medicare Advantage stocks trading down between 13% (Humana) and 25% (Clover). And this was all before the proposed rule bombshell on Friday, which seems to represent a massive overhaul of the program. More on that in the next section.

  • MA stewardship of taxpayer dollars. 2026 was an interesting year for the Medicare Advantage market. Overall, the sector outperformed as trends stabilized, resulting in beat-and-raise quarters throughout the year. At the same time, going back to the Advance Notice / Final Notice kerfuffle at the beginning of the year, CMS leadership made it clear it was unhappy with Medicare Advantage organizations’ stewardship of taxpayer dollars. Advance Notice suggested a rate reduction, and the industry pushed back hard, arguing the proposed rates would cripple the industry. Final Notice acquiesced, with rates coming up. Ultimately, the protest around Advance Notice seems set to backfire, as CMS has clearly still not forgotten the underlying point it was making about stewarding taxpayer dollars. The 2026 outperformance ultimately seems like a good way to poke the bear, as evidenced by CRUSH and the 2028 proposed rule. Perhaps Devoted was onto something last year in encouraging CMS to pay it less.

  • Encouraging competition… via MSSP? At the same time CMS is laying the hammer on Medicare Advantage, the MSSP program seems to be performing quite well. MSSP reported 2025 results this week, with Medicare generating $2.5 billion in savings relative to the benchmark. 80% of ACOs are earning performance payments, generating $4.9 billion for those ACOs. Recall that Klomp earlier this year shared CMS’s desire to enroll every Medicare beneficiary in an accountable care relationship. If you’re at CMS looking at both the current political climate and Medicare Advantage performance, it’s not hard to envision a scenario where MSSP and original Medicare look like an appealing way to compete with Medicare Advantage while you overhaul the MA program.

  • Devoted’s commercial shift. As regular readers know, I am fascinated by Devoted. After musing a few weeks back on their commercial market entry, I’ve heard from a handful of folks in the community who've shared their perspective that Devoted is indeed quite serious about the commercial market. Why is that relevant to bring up here? Devoted’s hard shift into commercial could be read as a bit of a tell on Devoted’s part about where the Medicare Advantage market is heading. There is inherent shareholder risk in being a single-line insurance plan in a stagnating market; cf. Oscar in the ACA. Perhaps Devoted is astutely attempting to get ahead of this risk now.

  • The 1957 baby boom peak. Speaking of Oscar, back in 2023 Mario Sclosser made an interesting point about MA market growth that I've been thinking about this week. 1957 represented the peak of the baby boom, which plateaued for several years. In 1962, births began declining steadily, reaching ~3.2 million per year by 1973. Why does this matter? Because people born in 1962 will turn 65 in 2027, we are at the start of a decade-long decline in age-ins.

My takeaway: Despite a reprieve in 2026 as the industry finally catches up to the trend in a post-v28 world, the Medicare Advantage market still faces several key headwinds in the years ahead, and that’s before you account for a 27% decline in people turning 65 over the next decade. The days of Medicare Advantage as the engine behind good vibes across the entire healthcare services sector seem behind us.

As we’ll get into more in the next section, CMS seems intent on ensuring the Medicare Advantage program is a good steward of taxpayer dollars, a level playing field for all plans, and focused on managing the needs of low-income/complex chronic populations. That’s still an important program, to be sure, although perhaps one that will be less exciting for investors.

Medicare Advantage, Part II

2. CMS goes all in on overhauling Medicare Advantage

By Wednesday, the Medicare Advantage industry had already had a bad week between Medicare for All and CRUSH. Then on Friday night, CMS dropped the proposed rule for 2028 Policy and Technical Changes to Medicare Advantage. It includes an overhaul of the current Stars program and a blueprint for Medicare Advantage's future, featuring inferred risk scores and a Diamond rating model that would replace Stars altogether, among other changes that appear designed to support low-income enrollees.

  • The proposed Stars overhaul. CMS is proposing an overhaul of measures included in Stars scores, resulting in fewer, harder-to-achieve metrics. For Stars nerds, other important changes include removing the reward factor and narrowing the improvement hold harmless to 4.5+ Star contracts only. As we’ll discuss more in a few bullet points, CMMI also appears to be testing the successor to Stars. This seems like big deal, assuming it is finalized.

  • A proposed $107 billion cut. CMS projects that, with these changes, 58% of Medicare Advantage contracts will see their ratings fall, with 20% losing their quality bonus payment. It expects this drop could generate ~$107 billion in savings over the next decade, modeled as follows.

  • CMMI’s proposed first mandatory model. CMMI is proposing a new mandatory model, ADVaNCE MA, starting in 2029 and running for ten years in twelve randomly selected states. It includes a number of elements that make it feel to me like CMS is laying its cards on the table that this is its proposal for what MA should look like in the future:

    • Pick one: CA, NY, FL, or TX. And eleven more. CMMI is bucketing states into seven categories and will randomly choose one or two states from each. The four states above (CA, NY, FL, TX) are Group Seven, and CMMI proposes selecting one state from this group to participate in this mandatory model. Ten states will come from Groups One through Five; one additional state will come from Group Six.

    • Inferred risk! Policy nerds have been expecting a move to an inferred risk model, given that CMMI Director Abe Sutton co-authored a piece with Gabe Drapos in Health Affairs proposing the concept in 2024. It appears that we’ll see it be shadow tested in ADVaNCE.

    • Stars turn into Diamonds. CMS leadership must have preschoolers with favorite shapes similar to mine, as it proposes a new Diamond Rating quality program that will replace Stars. The Diamond Rating appears to move to a more continuous scale, eliminating the 4-Star threshold issue, among other changes to the quality program.

My takeaway: CMS seems to be showing its hand in this proposed rule, with the ADVaNCE MA model in particular essentially laying out its vision for what Medicare Advantage should look like in the future, starting with a mandatory experiment in twelve states. It offers a fundamentally overhauled approach to risk adjustment via inferred risk, quality measurement via a new Diamond program, and lots of other adjustments that appear intended to both level the playing field among plans while also focusing on supporting complex and low-income populations.

Between CRUSH and this proposed rule, CMS seems intent on moving quickly to overhaul Medicare Advantage. There will be a lot to unpack in the weeks ahead. Smart folks are starting to share their perspectives on the news here, including our own Martin Cech (LI post), along with some of our favorite nerds Ali Khan (LI post) and Andrew Schwab (LI post).

It’s a big proposed change, but at this point, it's just that: a proposed change expected to start in 2029. Lots of time between now and then, including a 2028 presidential election. But it feels like CMS has taken its swing here.

Consumer Healthcare, Part III

3. LillyDirect and General Medicine launch Alzheimer’s partnership (and GLP-1s, too)

In Friday’s Daily newsletter, I dug into the expanding partnership between General Medicine and Lilly. Lilly invested in General Medicine’s $120 million Series B round announced last month. This week, Lilly and General Medicine announced two different partnerships. First, General Medicine is now available in LillyDirect as a telehealth provider for weight management. Second, General Medicine launched a memory care pathway, now one of three memory care options on Lilly’s website (alongside Isaac Health and Synapticure). As Ngai Yeung covered nicely in Endpoints, the goal is to support earlier screening and diagnosis of cognitive decline, leading to more prescriptions of Lilly’s Alzheimer’s drug Kisunla.

  • What’s the General Medicine store? In October, General Medicine raised $120 million to build an everyday healthcare store on its website. Then, later in October, it announced a partnership with Walmart to power the infrastructure layer for Walmart’s Better Care Services store. Now, it is a GLP-1 (and Alzheimer’s) telehealth vendor available on LillyDirect. I’m starting to lose the plot on what a storefront is in General Medicine’s parlance. It feels like a virtual medical group that offers a shoppable set of services. Perhaps that doesn’t matter all that much as they’re just focused on building demand, but I think it does, and TJ Parker has said as much as well. In the recent a16z conversation, he made the point that General Medicine is different from PillPack in part because it is a marketplace, whereas PillPack was the pharmacy. If the LillyDirect partnership is any indication, General Medicine isn’t yet a marketplace, it’s a medical group.

  • General Medicine’s Trustpilot reviews. I asked Claude to analyze the 1,058 General Medicine reviews on Trustpilot for key themes. The takeaways are below. The vast majority are positive reviews (957 are 5 Star reviews), primarily focused on speed and ease. A growing number of negative reviews exist on a small base, focused on billing issues and operational challenges. The issues with physicians not showing up on time for virtual appointments seem concerning, but perhaps also to be expected from a startup focused on growing demand first.

    • Note the GLP-1 comment from Claude as a particularly interesting one — people are switching to General Medicine for branded GLP-1s from other platforms.

      Claude’s summary of General Medicine Trustpilot reviews

  • The Alzheimer’s play. As I dove into in Friday’s newsletter, I find the Alzheimer’s piece to this partnership much more interesting strategically than the GLP-1 component, which seems straightforward. Lilly’s drug Kisunla is a once-monthly infusion for mild cognitive impairment. This creates a challenge for Lilly in that it needs earlier screening to detect mild cognitive impairment, as many people are past this stage by the time they reach a specialist. Lilly needs to get more people screened earlier to drive sales of Kisunla (I noted in Friday’s NL this gives me flashbacks of learning about how the PHQ-9 came into existence to support SSRI sales). Now, enter a theoretically abundant screening source into the equation: AI-centric virtual care. If General Medicine can help people screen for cognitive impairment earlier, it should then get more people on Kisunla, which Endpoints notes is the definition of success in this partnership.

  • Pharm-to-table gets complex. If this Lilly / General Medicine partnership is successful and indeed drives meaningful new sales of Kisunla, it seems to call into question one of the core critiques I’ve heard from the industry about D2C-style online models, which is that they’re only for consumer-y needs, and that people are still going to see doctors for real serious healthcare issues that require care coordination. I’m not so sure that argument will hold, and this partnership points to why. It is incredibly easy to sketch out a scenario where someone worried about their brain health goes to their personal AI agent, asks a few questions, and the AI agent serves up this General Medicine care pathway as an option. A same day video visit to check my brain health? Great! General Medicine and Lilly will then need to figure out the integration into care delivery from here — Kisunla appears to require a confirmation of amyloid buildup for diagnosis, and then is delivered via a once-monthly infusion. How General Medicine manages the referrals out for those pieces of care is going to be fascinating.

    • In conversations with health system leaders, one of the more worrisome “what if” scenarios revolves around whether physicians will become viewed as impediments for consumers who are seeking treatments they’ve decided they want. This partnership presents a very real scenario of how that change can happen. It reminds me of the frustrated NEJM Voices perspective from a cardiologist in September, lamenting how a patient reached out for a test she didn’t think was necessary. While physicians might be frustrated with this type of scenario, it’s easy to see how consumers will prefer the ease and cost of the D2C style experience.

  • Kisunla’s black box warning & headline risk. I am by no means the world’s expert on all things FDA, but as I understand it, Kisunla has a black box warning for causing potentially fatal brain bleeds. A black box warning is the FDA’s most serious safety warning. There are already exposes online about the concerns people have with this new class of Alzheimer’s drugs. If this care pathway takes off for General Medicine as planned, the math indicates something bad will happen at some point. In Kisunla’s Phase III trial, 4 out of 853 patients on the drug had a large brain bleed, versus 2 of 874 on placebo. In today’s world, that makes for easy fodder for an investigative journalist, particularly when Lilly has invested in the medical group that is pushing the screening for the drug. From where I sit, I wonder if Lilly is appropriately assessing the headline risk of owning a stake in a virtual medical group prescribing this drug.

My takeaway: This Alzheimer’s pathway feels like an interesting test case for how far consumers will go in directing their own complex health journeys, and also what the blowback will look like politically as negative outcomes inevitably arise. General Medicine’s Trustpilot shows that while it’s doing a lot right, it is still a startup with the scaling issues one might expect. Now, it’s entering a partnership with Lilly to diagnose patients and get them into treatment faster for a black-box-warning drug. It is easy to see why consumers will start going down this path; it is equally as easy to see the coming political backlash.

Either way, if I’m a health system leader thinking about the future of care delivery, I am keeping a close eye on this play and probably investing heavily in high-end infusion clinics to deliver these medicines.

Heard Among Nerds

4. Oura/Hims/Ro and the battle for D2C supremacy

Here are the results of last week’s poll asking which consumer brand y’all felt is best positioned for the most durable success in the years ahead. Oura won by an overwhelming margin.

Competition in this D2C space is evolving rapidly, as evidenced by the section above on General Medicine and Claude noticing that it is winning in GLP-1s based on Trustpilot reviews.

Further evidence of this came on my drive to the coffee shop to finish writing this morning; I heard more radio ads from OrderlyMeds offering me low-cost access to GLP-1s. When I Google “OrderlyMeds”, I see sponsored results from Ro, AltRx, Mochi Health, Refills.com, and more, all competing for the right to sell GLP-1s (whether branded or compounded). Given Hims’ membership decline last quarter, they may have a real problem on their hands here as various entities spend on acquiring GLP-1 members. It all seems reminiscent of the virtual therapist market years ago, when acquisition costs skyrocketed, and massive churn eroded lifetime value for companies investing in the market.

The Hims peptide launch may offer another interesting data point. Folks following the Hims and broader longevity story know the much-anticipated peptide launch is coming, as Hims opened a waitlist for its Peptide+ product a few weeks ago. Hims’ CEO Andrew Dudum took to X to share that the waitlist reached 10,000 people as of Thursday. While that number seems, at first glance, to be designed to excite the market, I can’t help but wonder why it wasn’t higher for a company with 2.8 million customers that is seemingly betting the farm on this launch. For comparison, Neko attracted a 25,000-person waitlist of New Yorkers for its new NYC clinic launch.

Primary Care

5. Boston’s primary care dilemma

As regular newsletter readers know, I’ve been fascinated by the state of primary care in Boston as a microcosm of the perplexing state of US healthcare. Boston is a place where people theoretically have access to the best medical care in the world, but in some cases can't get a primary care visit for over a year. The juxtaposition of those two concepts is befuddling.

  • Revisiting MGB’s issues. Back in May, I dove into a proposed partnership between CVS and MGB, which was designed to increase access to primary care services in the market, though regulators worried about cost increases. KFF highlighted earlier this year the insane waitlists patients face in Boston to access care, and that MGB was also working with K Health to offer AI-centric primary care. MGB has received significant blowback over this, as its PCPs have voted to unionize and just this week sent letters to state officials urging MGB to invest more in primary care.

  • The Boston Globe series. The Boston Globe published a series this week exploring the myriad challenges facing primary care in the market, including a survey of 253 PCPs in Boston expressing their perspectives on why primary care is broken. It doesn’t strike me as coincidental that MGB’s PCPs were also writing letters to state officials this week, suggesting Boston-area PCPs are making a strong push to protect their profession.

  • Numbers that confuse me. I must admit, I always get more confused when I wade into “doctor shortage” stats. According to Healthline, Boston has 1,384 primary care doctors. According to KFF, there are 17,979 professionally active PCPs in Massachusetts. The Boston Globe survey suggests the average PCP sees ~1,500 patients. Multiplying that by 17,979 professionally active PCPs suggests Massachusetts’ PCPs should have the capacity to see 27 million people. Massachusetts population is roughly 7.2 million people. The math obviously overstates things, among other reasons, because KFF has a broad definition, but even a massive haircut to that math doesn’t get you to a dire shortage.

  • This issue isn’t evenly distributed. The Boston Globe experience, along with previous KFF articles, makes it clear that regardless of what the shortage numbers say, people clearly have issues accessing doctors in Boston. That’s what matters. One thing that comes across from the Boston Globe’s experience in scheduling visits is that this issue isn’t consistent across organizations. A scheduler for Mass General Hospital said there was no availability and to call back in December (this was an August call). Cambridge Health Alliance and Tufts both had appointments within a week, making it clear that it’s not everywhere in Boston that this issue exists. The data below from the article further supports that conclusion, as independent practices have significantly more availability:

  • Real challenges; unclear solutions. As the article notes, primary care doctors face myriad issues. A retirement cliff (20% are 65+). High cost of living in Massachusetts. Issues in medical schools. Market consolidation. Doctors aren’t seeing patients full-time. Insurance reimbursement is a challenge. But overwhelmingly, 84% of Boston PCPs blame the institutions and medical schools in the market for prioritizing specialty care over primary care.

  • Why should MGB invest more in primary care? If I put myself in a regulator's shoes, I’m very much on board with spending more money on primary care in Massachusetts. But for the life of me I don’t understand why I should encourage MGB to invest in primary care. MGB very clearly does not want to! The evidence of this is extensive at this point: 1. PCPs are so unhappy there that they’ve unionized. 2. MGB has claimed to commit $400 million to primary care, but the PCPs aren’t seeing it. 3. It is impossible to access primary care at MGB, when Tufts and others have solved this issue. 4. It is prioritizing partnerships with CVS and K Health to offer primary care instead. 5. It is financially distressed as a system already. I’m sure primary care leadership at MGB would see this as a leadership issue at MGB and encourage an overhaul there. But that hardly strikes me as a regulator's role. If MGB wants to focus on more specialty care and offer proton beam therapy, it should do that. But I’d actually encourage MGB to rip the bandaid at this point. Clarify its strategy and exit primary care entirely, moving to a partnered approach where it encourages other organizations to provide primary care.

  • PCPs are fighting for the future of their profession. If one thing is clear from this series of articles, combined with the MGB case study, it’s that local primary care providers are actively fighting to protect the future of their profession. What MGB is doing actually strikes me as a logical strategic decision for an academic system: partner with institutions like CVS and K Health to expand access to primary care. AI and retail care options should be available as options in a market for consumers to choose from.

  • If I’m General Medicine, I am putting billboards all over Boston yesterday. Why not? Seemingly, access is a huge issue in the market, especially those in MGB’s orbit. Look at the Trustpilot reviews above. General Medicine is winning for patients in care gaps. That sounds like all of Boston based on these articles! I’d imagine every D2C online company will be looking at Boston saying, hey, we can help here! And frankly, that seems like a winning mindset. This seems like a huge entrepreneurial opportunity for organizations seeking to open up access in a market.

My takeaway: The problems in Boston with primary care access seem like a particularly acute case example of the tensions more broadly facing care delivery across the country. Yet I can’t help but notice the difference between the problem and proposed solutions in articles like the Boston Globe. The problem is generally that patients can’t access care, while proposed solutions focus on better supporting PCPs, with the presumption being that solving PCPs’ problems is what patients ultimately need. This is where the issue lies, as AI is increasingly breaking the premise that patients need to rely on PCPs as their agents for accessing healthcare, presenting retail and AI-centric solutions as viable alternatives. Whether or not that is a good thing feels like a good debate to have over Thanksgiving, but it certainly seems like it is a thing.

If I’m a regulator in Massachusetts, I’m juxtaposing the conversations about General Medicine and Boston and thinking about how to solve the access problems in the state, which will inevitably involve some combination of all of the above. And if I’m a PCP in Massachusetts, I’m thinking about how to embrace these shifts and move out of academic institutions and toward models more like Liza Health (a Boston-based primary care startup with some smart folks behind it).

Other headlines from the week:

  • MSSP generated $2.5b in net savings for Medicare in PY25 and paid out $4.9b in shared savings to ACOs, per CMS. Gross per capita savings rose 7.3% to $690. Net per capita savings slipped 2.9% to $234 as more ACOs moved into two-sided risk. Participation is growing again for PY26, with assigned beneficiaries up ~12.5%. The usual benchmark caveats apply, but it's a good backdrop for CMS's MSSP-friendly posture in section 1.

  • CVS Health published a white paper using Aetna commercial claims. It estimates GLP-1s at current prices raise total cost of care by ~$3,900 per member by year three, with breakeven around ~$350 a month. CVS is hardly a neutral party in the GLP-1 pricing fight, but its numbers land in a similar range to a Lilly study.

  • Also on GLP-1s, Blue Shield of California is pitching a monthly "subscription fee" to manufacturers as an alternative way to pay for GLP-1s, plus bonuses when patients hit outcomes like lower blood sugar or weight loss.

  • NewYork-Presbyterian is acquiring Westmed Medical Group from Summit Health as Sycamore continues to sell VillageMD assets piece by piece. The deal also includes Summit's New York locations in Manhattan, Brooklyn, Queens, and Long Island. Westmed had nearly 500 physicians across 13 locations when Summit acquired it in 2022.

  • A federal judge dismissed Tampa General's lawsuit against Eli Lilly over Lilly's requirement that 340B hospitals submit prescription-level claims data. The judge held that Tampa General can't use Florida consumer protection law to challenge the requirement and pointed instead to HHS's administrative dispute process.

  • Nebraska will skip CMS's optional six-month "hold harmless" period for Medicaid work requirements, which runs through June 30, 2027 for states that opt in. Nebraska adopted the requirements early, on May 1, the first state to do so. Nebraska Appleseed says more than 1,000 Nebraskans have been denied care or lost coverage since May, based on two-month-old data.

  • Humana is expanding its gold card program nationally. About 50k providers will be exempt from prior auth for MRI, CT, and MRA imaging, plus physical and occupational therapy.

  • Molina sued CMS over a $140m MA Star Ratings loss. It joins ~13 other plans, including Alignment Health, CVS Health, and Elevance.

  • Menlo Ventures' State of AI in Healthcare report estimates healthcare AI spend grew 2.5x to $3.7b in 2026. That makes healthcare the largest vertical AI market, more than double legal at $1.8b.

  • OpenEvidence reported it has surpassed two million consults a day, roughly one for every two U.S. office visits. My napkin-math diffusion model suggests a plateau around ~3.2m a day towards the end of next year. OpenEvidence also got a notable shout-out: Daniela Amodei mentioned Anthropic’s partnership with OpenEvidence twice in her brief remarks as Anthropic's IPO priced.

  • Hone Health, a longevity-focused telehealth clinic, raised $228m. That includes $200m in growth financing from General Catalyst's Customer Value Fund and a Series A extension led by MarcyPen Capital Partners. Recall that GC’s Customer Value Fund is a debt instrument designed to finance upfront CAC costs.

  • Solarte Health raised a $3m seed to connect independent doctors with self-funded employers. It's focused on Minnesota and Wisconsin employers with 50 to 3,000 employees, with USI Insurance Services as a brokerage partner and Nice Healthcare as a preferred DPC partner.

Want insights like these seven days a week?

HTN Pro membership gets you the Daily Brief newsletter in your inbox Tuesday through Friday mornings; access to our conversations with industry experts; and membership to the HTN Slack community. Learn more here.

Want to comment or share feedback?

Login or Subscribe to participate

If this newsletter was forwarded to you, subscribe here or see more from Health Tech Nerds.