Read more

The Number CMS Leads With, and the One It Doesn't

When CMS announced the 2027 Medicare Advantage landscape in early October, its official messaging emphasized stability. More than 99 percent of beneficiaries have access to at least one MA plan. About 8 in 10 enrollees can keep their current plan at the same or lower premium. The average monthly MA premium is projected to fall 16.5 percent, from $14.37 to $12 a month.

All of that is technically accurate. None of it captures what is happening at the county level.

A KFF analysis published October 9 using CMS 2027 landscape data found that 193 counties will have no Medicare Advantage prescription drug plan available for general enrollment. In 2026, that number was 123. The one-year increase is 70 counties -- a 57 percent jump in a single plan year.

About 638,000 beneficiaries live in those 193 counties. In 2026, 399,000 people were in that position. The gap widened by 239,000 people in one year.

The 99 percent figure and the 57 percent figure coexist. They are not contradictory. What they measure is different. CMS counts whether a plan exists somewhere in a county. KFF counts whether a plan with prescription drug coverage exists for people who aren't in a special needs population. Those two measurements answer different questions, and the one that matters for daily medication access is the second one.

The counties joining the no-plan list in 2027 are in 16 states plus U.S. territories other than Puerto Rico: Alaska, California, Colorado, Idaho, Kansas, Minnesota, Montana, Nebraska, Nevada, New Hampshire, North Dakota, Oklahoma, Oregon, South Dakota, Utah, and Vermont. Three states -- New Hampshire, North Dakota, and Oklahoma -- are newly affected compared with 2026.

Alaska is the sharpest single-state example. No county in Alaska offers a general-enrollment Medicare Advantage prescription drug plan in 2027. That affects about 110,000 beneficiaries. Six counties in Nebraska and South Dakota have a Medicare Advantage plan without drug coverage, meaning residents can enroll in an MA plan but must buy a separate Part D plan for their medications. The other 187 counties in the no-plan list have nothing -- no MA plan, with or without drug coverage.

The Insurer Exit Map, by the Numbers

The county-level picture is the product of decisions made by individual insurers. The KFF analysis tracks county entries and exits for each company. The pattern is not uniform.

Health Care Service Corporation -- which owns Blue Cross Blue Shield plans in Illinois, Montana, New Mexico, Oklahoma, and Texas -- is exiting 498 counties and entering zero. That is the largest single-insurer reduction of any major carrier. HCSC acquired Cigna's Medicare business in 2025. In 2027, its Blue plan footprints are contracting sharply: Texas fell from 229 counties to 47, Illinois from 100 counties to 7, Oklahoma from 58 counties to 6. The company is exiting Connecticut, Kansas, Montana, New York, Oregon, Utah, and Washington D.C. entirely.

Centene is exiting 359 counties and entering 15. The company is leaving Hawaii, Oklahoma, and Tennessee, and moving to SNP-only status in Nevada and Ohio. Its total plan count falls from 303 to 210 -- a reduction of nearly one-third.

UnitedHealth Group is exiting 228 counties and entering 9. UHG's presence falls from nearly 80 percent of U.S. counties in 2026 to about 70 percent in 2027. The company is pulling standard plans from 140 counties and exiting 34 counties entirely, including all of Wyoming.

Humana is exiting 62 counties and entering 4. It is leaving Minnesota entirely. Its plan exits affect approximately 600,000 members -- notable context given that it added more than 1.2 million members during the previous enrollment period.

CVS Health's Aetna business is exiting 186 counties and entering 63. Aetna is pulling standard plans from 123 counties, exiting Maryland and New Hampshire, and is down from 2,139 counties to 2,072.

Blue Cross Blue Shield of Michigan Mutual is exiting 177 counties and entering 1. It will offer Medicare Advantage prescription drug plans only in Michigan and Wyoming, down from five states.

Group 1001 -- which operates through Clear Spring Health -- and Molina Healthcare are exiting the general Medicare Advantage prescription drug market entirely. Molina will continue to offer special needs plans in many counties, but its general-enrollment MA-PD presence ends.

In total, 12 payers offering Medicare Advantage in 2026 will not offer it in 2027, eight of them health system-owned plans (including Providence, Memorial Hermann, and Christus Health). Wellmark, BCBS North Dakota, and BCBS Mississippi are also among the exits.

On the other side: Devoted Health is expanding, entering 342 counties with no exits -- the largest expansion of any insurer in the 2027 landscape. Devoted is adding chronic condition plans in 30-plus states and more than doubling its C-SNP count from 122 to 255. Four organizations are entering the MA market for the first time in 2027, including St. Luke's Health Plan in Idaho and MultiCare Health System in Washington state.

The national plan count moves from 5,553 to roughly 5,532 -- a minimal net change. That figure is the headline. The geographic churn behind it is the story.

What's Driving the Exits: The Payment Math

The proximate cause is the same across nearly every exit announcement: CMS payment rates are not covering costs.

CMS published its final 2027 payment rule in April 2026, setting a 2.48 percent rate increase for Medicare Advantage. That increase equals approximately $13 billion in additional plan payments. When risk score trends are included, the effective rate increase is closer to 4.98 percent.

Insurers and provider groups said it wasn't enough.

The American Medical Group Association stated that the update "fails to keep pace" with the costs of delivering Medicare care. AMGA cited 1.8 million MA enrollees in 2024 whose plans were unavailable in 2025 as evidence of a structural underpayment pattern. Industry groups had pushed back hard on the original proposal -- CMS initially proposed a 0.09 percent increase, then raised it to 2.48 percent after comment.

UnitedHealthcare President Bobby Hunter said in a September 2026 briefing that MA funding "has not kept pace with medical and pharmacy costs." He cited "funding pressures, rising medical costs, rising drug costs, and increased utilization" as drivers of the company's county pullbacks. He also noted that "more than 95% of our non-special needs plan members never reach" the out-of-pocket limit -- a reference to the cap introduced under the Inflation Reduction Act, which limited what beneficiaries pay in a year but shifted costs onto plans.

The Inflation Reduction Act's Part D redesign changed the financial architecture of Medicare drug coverage. Before the redesign, a portion of catastrophic drug costs was covered by the federal government's reinsurance program. After the redesign, more of those costs shifted to plans directly. For carriers with a high proportion of members on expensive medications, that shift changed the underlying economics of the product.

The V28 risk adjustment changes have also pressured plans. CMS has been transitioning to a new risk adjustment model (V28) that generates lower risk scores for many beneficiaries, reducing per-member payments to plans. CMS delayed further changes to the V28 approach for 2027 to let the existing changes "settle," but the delay is described as indefinite, not permanent.

The result is a market in which the carriers best able to absorb cost pressure are maintaining or expanding their footprints. Those with thinner margins or less geographic diversification are pulling back. HCSC's Texas compression -- from 229 counties to 47 -- is a visible version of that pattern. CVS/Aetna is entering 63 counties while exiting 186; it is not retreating uniformly, but repricing where it can and withdrawing where it can't.

The Part D Standalone Collapse

The counties with no Medicare Advantage prescription drug plan are not necessarily without any options. Beneficiaries can use traditional Medicare and buy a standalone Part D prescription drug plan. But the standalone Part D market is also contracting.

KFF found that the average number of standalone Part D plans available per beneficiary falls from 11 in 2026 to 9 in 2027. The total plan count drops from 360 to 316 nationally.

The more significant change: no zero-premium standalone Part D plan will be available in 2027 to people who don't qualify for low-income assistance. In 2026, some zero-premium standalone Part D plans existed. They are gone.

Approximately 4 million Part D enrollees who paid no premium in 2026 will pay one in 2027, regardless of whether they stay or switch plans. KFF found that AARP Medicare Rx Preferred -- a widely enrolled plan -- is raising premiums by $50 or more per month in every state. The CMS estimate of the average standalone Part D premium moves from approximately $35 to $36 monthly. That average includes the cheapest plans: Wellcare Value Script at $5.30 a month in 45 states and Wellcare Classic at $7.30 in 33 states. The average will be higher for most people in practice because the cheapest plans limit their formularies to keep premiums low.

New York's standalone Part D floor is $37.30 per month -- seven times the national floor. In Wyoming, the highest average MA premium of any state is $69.23. These are not national averages. They are what specific beneficiaries in specific geographies will actually pay.

The Part D stabilization program that CMS operated during the IRA transition is ending after 2026. The program subsidized premiums during the phase-in. Its expiration means the full cost of the IRA's Part D redesign flows through to enrollees and plans beginning in 2027.

What Beneficiaries Actually Have Left

The loss of a Medicare Advantage plan does not mean the loss of Medicare coverage. Every beneficiary in a county with no MA-PD for general enrollment retains access to traditional Medicare -- Parts A and B. They can buy a standalone Part D prescription drug plan. They can, if they choose, also purchase a Medigap supplemental policy.

The enrollment protections for people in exiting markets are meaningful. Enrollees whose MA plan leaves their area get a special enrollment period that can extend past the December 7 open enrollment deadline. Those switching from MA to traditional Medicare may have a time-limited guaranteed-issue right to buy certain Medigap policies without medical underwriting. That right is not unlimited and varies by state -- some states have protections that extend it, many do not.

The practical barrier is not legal. It is logistical and financial.

A beneficiary in a rural county who loses their MA plan must navigate the transition during a period (October 15 to December 7) while still managing their health. They need to confirm that their medications are covered by whatever Part D plan they select. They need to understand whether their physicians and specialists are in-network under any remaining MA option or, if switching to traditional Medicare, to verify that their providers accept Medicare assignment. They may need to purchase a Medigap policy -- whose monthly premium is separate from Part D -- to limit their exposure to traditional Medicare's cost-sharing.

CMS Administrator Dr. Mehmet Oz urged beneficiaries to use open enrollment as "an opportunity to review their coverage." The 1-800-MEDICARE hotline and the Medicare Plan Finder tool are available resources. Each state's State Health Insurance Assistance Program offers free, unbiased counseling. These resources exist and are genuinely helpful. They do not eliminate the burden of transition for people who are already managing chronic conditions or limited transportation.

There is also the enrollment cap problem. Fifteen insurers have enrollment caps on 65 plans across 11 states in 2027, with caps ranging from 410 to 56,345 members. Capped plans are closed to new enrollees until membership falls. A plan that technically exists in a county but refuses new enrollment does not function as an option for someone whose prior plan exited.

The Rural Pattern and Why It Persists

The geography of the exits is not random. Rural counties absorb disproportionate reductions because the economics of rural MA delivery are structurally different from urban markets.

Urban MA markets feature higher enrollment concentration, more in-network providers competing for contracts, and shorter travel distances that make specialist care tractable within an HMO model. Rural markets feature lower enrollment, fewer competing providers, and geographic spread that makes network construction expensive and narrow-network plan design impractical.

The four Minnesota counties that lose all three of their Medicare Advantage insurers simultaneously -- Brown, Lake of the Woods, Morrison, and Renville -- are a concrete version of this pattern. Three separate insurers, operating under independent business decisions, reached the same conclusion about the same four counties in the same year. That is not a coincidence. It is the market's pricing of rural risk.

Health Care Service Corporation leaving Miami-Dade County and Humana exiting Monroe County, New York -- which includes Rochester -- shows that urban markets are not immune. But the magnitude of urban exits is bounded by competing plans that fill the space. In Miami-Dade, other carriers remain. In rural Brown County, Minnesota, they don't.

The 28-plan national average obscures this. KFF notes that the national average is heavily shaped by urban markets. Lancaster County, Pennsylvania, has 66 Medicare Advantage prescription drug plans available. That figure pulls up a national average that includes zero-plan counties. The median -- what a typical rural beneficiary experiences -- is lower than the mean.

The relationship between the MA market and rural healthcare access is not new. The National Rural Health Association has tracked Medicare Advantage penetration in rural areas for years, noting that rural beneficiaries have historically had fewer options and have relied more heavily on traditional Medicare. The 2027 exits are an acceleration of an existing pattern, not a new one.

The 193-county figure will likely rise. The underlying conditions -- payment rate pressure, risk adjustment changes, the IRA's Part D redesign costs flowing to plans -- are not resolved for 2028. Insurers that are currently trimming to profitable geographies will continue to assess where their margins hold. Counties that currently have one or two options are at higher risk of joining the no-plan list in future years than counties with ten.

What This Means for Lung Cancer Screening

Medicare Advantage plans and lung cancer screening have a specific relationship that makes the exit map more than an abstract policy problem for Oatmeal Health.

The U.S. Preventive Services Task Force recommends annual low-dose CT screening for adults aged 50 to 80 with a significant smoking history. Medicare -- both traditional and MA -- covers the scan at no cost-sharing for eligible beneficiaries. The coverage rule is the same in both programs. The access gap is not in the coverage rule. It is in the counseling, the navigation, and the follow-through.

Medicare Advantage plans are required to cover a shared decision-making visit before a lung cancer screening can be ordered for most beneficiaries. That visit is also covered under traditional Medicare. But in practice, the visit completion rate is low. Research has found that the counseling conversation that unlocks access to screening happens at well under 20 percent of the rate it should for eligible patients.

The communities where MA exits are concentrating -- rural, lower-income, with thinner healthcare infrastructure -- are also the communities where lung cancer incidence and mortality rates are highest, where primary care access is most constrained, and where the gap between eligible patients and screened patients is widest.

A beneficiary in a no-plan county who transitions to traditional Medicare does not lose their legal right to a low-dose CT scan. They lose the care coordination infrastructure, however limited, that an MA plan may have provided -- the reminder letters, the care management programs, the navigator calls. In rural markets where Oatmeal Health operates, that infrastructure was already thin. Its further erosion through plan exits is a meaningful headwind for screening access.

Oatmeal Health's approach to this problem is not to replace the insurance infrastructure. It is to operate at the community level -- partnering with rural health systems, community health centers, and primary care practices to identify high-risk patients and connect them to screening. The MA exit map is a background condition, not a barrier to that work. But it is a reminder that the coverage map and the access map are not the same thing, and in rural America the gap between them tends to be widest precisely where the need is greatest.

About

Jonathan Govette is Co-Founder and CEO of Oatmeal Health, a company focused on expanding lung cancer screening access for underserved populations. Oatmeal Health partners with health systems, community health centers, and primary care practices to identify high-risk patients and connect them to low-dose CT screening. Learn more at oatmealhealth.com.

Key References

• KFF. "Medicare Advantage Plan Availability Varies Widely by County for 2027." October 9, 2026. https://www.kff.org/medicare/medicare-advantage-plan-availability-varies-widely-by-county-for-2027/

• Becker's Hospital Review / Becker's Payer. "Medicare Advantage Looks Stable for 2027. Underneath, It's Shifting." October 2026. https://www.beckershospitalreview.com/finance/medicare-advantage-looks-stable-for-2027-underneath-its-shifting/

• Fierce Healthcare. "CMS Gives Medicare Advantage Rates a 2.48% Bump for 2027 Plan Year in Final Rule." April 2026. https://www.fiercehealthcare.com/regulatory/cms-gives-medicare-advantage-rates-248-bump-2027-plan-year-final-rule

• Medical Daily. "638,000 Medicare Beneficiaries in 193 Counties Will Have No General Medicare Advantage Drug Plan for 2027." October 2026. https://www.medicaldaily.com/medicare-advantage-no-plan-193-counties-2027-kff-479856

• Becker's Hospital Review. "Humana to Exit Medicare Advantage Plans Covering 600,000 Members in 2027." 2026. https://www.beckershospitalreview.com/finance/humana-to-exit-medicare-advantage-plans-covering-600000-members-in-2027/

• Goodsurance. "KFF: Number of Counties With No Medicare Advantage Plan Rises 57 Percent for 2027, With UnitedHealth Group Exiting 228 Counties." October 9, 2026. https://www.goodsurance.com/news/2026-10-09-kff-ma-county-availability-2027-rural-uhc-humana-exits

• Oatmeal Health. "Medicare Advantage Mass Exit 2027." October 17, 2026. https://oatmealhealth.com/medicare-advantage-mass-exit-2027/

• CMS. "2027 Medicare Advantage and Part D Final Rule." April 2026.

Read more

A papal diagnosis brought lung cancer back into the headlines. The data behind the screening gap has been there all along.

On October 13, 2026, a team of surgeons at Rome's Gemelli Hospital removed the lower lobe of Pope Leo XIV's right lung. The Vatican confirmed what pre-operative tests had suggested: pulmonary adenocarcinoma. The procedure was described as successful. No complications. His post-operative course, the Vatican said, is normal.

Within hours, lung cancer was trending. Search volume spiked. Clinics reported phone calls from people asking about screening. Coverage ran in every major outlet. Then, within a few days, the story faded into the next news cycle.

That is the predictable arc of celebrity diagnosis coverage. A famous person gets sick. The public briefly notices. Nothing structural changes.

Here is what did not make the news cycle: the American Cancer Society had already published a rigorous model showing that full screening of all eligible Americans could prevent roughly 62,110 lung cancer deaths over five years. At current rates, we prevent about 14,970 of those. That leaves approximately 47,140 preventable deaths in the gap, every five years, not because the tool does not work, but because most people who should get it never do.

Pope Leo's diagnosis did not create that gap. It just made it visible, briefly, to people who do not spend their days reading radiology journals.

This piece is for the ones who do, and for the health system operators, FQHC leaders, radiology administrators, and benefits directors who can do something about it. We will walk through what the Vatican actually disclosed, why two credible surveys produce two different screening rates, what stage at diagnosis does to survival, who the system keeps missing, what a working pathway looks like in practice, and where technology fits honestly into that picture.

1. What the Vatican said, and what it did not

The facts released publicly are narrow, and they were released slowly.

The nodule was found during routine medical checks, according to the pope's own statement. Surgeons at Gemelli, led by thoracic surgeon Dr. Stefano Margaritora, performed a video-assisted resection targeting what they called a radical removal of the right lower lobe. Intraoperative examination confirmed adenocarcinoma. The surgery lasted roughly 90 minutes.

What has not been disclosed: the stage of the cancer, whether lymph nodes were involved, what imaging method first detected the nodule, and his smoking history. Some early reports speculated about the stage based on the resection approach. Those speculations remain unconfirmed by the Vatican.

That ambiguity matters for one reason: it is tempting to make Pope Leo's diagnosis into a clean narrative about the power of early detection. Maybe his cancer was found early because of a well-designed surveillance program. Maybe it was incidental. We do not know. And turning an unconfirmed story into a public health parable has a long history of backfiring when the facts come out differently.

What the case does do is create an opening. When a head of state announces a lung cancer diagnosis, the general public becomes briefly receptive to information about lung cancer. People who have never thought about screening eligibility ask their doctors. Clinics get calls. The question for health system operators is whether the pathway is ready to absorb that attention, or whether the interest evaporates because nothing actionable is easy to find.

Right now, for most health systems, the pathway is not ready.

2. Two surveys, two numbers, one persistent problem

Search for the national lung cancer screening rate and you will find 18.7%. You will also find 24.5%. Both numbers are real. They come from different surveys with different methodologies, and understanding the gap between them is more useful than picking one.

The 18.7% figure comes from an American Cancer Society analysis published as a research letter in JAMA in November 2025. The researchers used the 2024 National Health Interview Survey. Among adults ages 50 to 80 who had ever smoked, 2,124 respondents met 2021 USPSTF eligibility criteria. Weighted to the national population, that represents about 12.76 million Americans. Of those, 18.7% were up to date on lung cancer screening. Priti Bandi, Ph.D., a senior principal scientist at ACS, summarized the finding plainly: roughly three times more deaths could be prevented if everyone eligible were screened.

The 24.5% figure comes from a research letter published April 27, 2026, in JAMA Internal Medicine. Todd Burus, Ph.D., and colleagues at the University of Kentucky used the 2024 Behavioral Risk Factor Surveillance System, a larger state-based telephone survey. Among 26,104 eligible respondents, 24.5% had been screened within the past year, a gain of about six percentage points from 2022 BRFSS data.

Why the gap between the two numbers? Different survey architectures produce different response patterns. Different questions get different answers. The NHIS analysis had a 47.9% response rate, which itself introduces uncertainty. Neither survey is wrong. They bracket a range. Somewhere between roughly one in five and one in four eligible Americans is actually getting screened.

For a health system leader, the useful reading is not which number to cite. It is that even the better number means three out of four eligible patients in your catchment area are not getting a scan that is fully covered, without cost sharing, by Medicare and by most ACA-compliant private plans. The clinical tool is free to patients who qualify for it. Three in four still do not get it.

One comparison sharpens the picture: a separate JAMA study examined 11,147 adults who were eligible for both breast cancer screening and lung cancer screening. About 65% had a mammogram in the prior year. About 18% had low-dose CT for lung cancer. Same population. Two different tools. One has achieved broad uptake. The other has not.

3. The gap in lives: what the model says

The ACS analysis in JAMA went beyond measuring uptake. It modeled what lung cancer screening could do over five years if it actually reached eligible adults.

The model drew on the National Lung Screening Trial, which found that annual low-dose CT reduced lung cancer mortality by 20.4% compared with chest X-ray. That figure served as the effect estimate. Applied to the roughly 12.76 million eligible Americans, the analysis projected:

At 100% uptake: approximately 62,110 deaths prevented and 872,270 life-years gained over five years.

At current uptake (18.7%): approximately 14,970 deaths prevented and 190,030 life-years gained over five years.

The arithmetic is direct. Current screening is delivering roughly 24% of what full screening could deliver. Subtract 14,970 from 62,110 and you get 47,140 preventable deaths not being prevented. That is the number behind this article's headline.

To be precise about what this figure is and is not: it is a modeled estimate grounded in trial data, not a registry count. Effect sizes from clinical trials do not always translate to routine practice at the same magnitude. A VA Health System study published alongside the Kentucky letter found higher all-cause mortality among screened veterans compared with National Lung Screening Trial participants, suggesting that real-world benefit may be smaller than trial benefit. The authors of that study do not argue against screening. They argue for quality: standardized reporting, competent reading, and robust follow-up. The gap between trial efficacy and real-world effectiveness is a reason to invest in pathway quality, not a reason to screen less.

But even if the real-world effect is half what the trial showed, the preventable death toll from low uptake is still enormous. The math does not become comfortable at any plausible discount.

The ACS analysis also modeled screening among the 28.08 million adults ages 50 to 80 who have smoked but do not meet current USPSTF eligibility criteria. Including them could, in theory, prevent an additional 29,690 deaths and gain 482,410 life-years. This raises harder questions about harms, false positives, and cost-effectiveness at the margin. Multiple oncology organizations have urged caution before expanding criteria ahead of trial evidence. We address that debate in section 5.

For now, the near-term arithmetic is clear. The biggest available gain is among the people who already qualify and are not getting scanned.

4. Stage is the whole game

The reason screening matters comes down to one table in the SEER database.

Among lung cancers diagnosed between 2016 and 2022, about 24% were found at a localized stage. Five-year relative survival for localized lung cancer is 65.5%. About 21% were diagnosed at the regional stage, with 38.2% survival. About 51%, half of all diagnosed cases, were found at the distant stage, meaning the cancer had spread to other organs or distant lymph nodes. Five-year relative survival for distant-stage lung cancer is 10.5%.

Read those numbers one more time. Half of all lung cancers are found after they have spread to distant sites. For those patients, roughly nine in ten will not be alive in five years. For the quarter caught while localized, about two in three will be.

The argument for early detection is not theoretical. The survival differential between stage I and stage IV lung cancer is one of the largest in oncology. Treatment has improved substantially. The American Cancer Society's 2026 statistics report shows overall five-year lung cancer survival rising from 15% in the mid-1990s to 28% now. Distant-stage survival rose from 2% to 10%. Regional-stage survival rose from 20% to 37%. These are meaningful gains from better therapeutics, including targeted agents and immunotherapy.

But the single largest lever, shifting diagnosis from distant to localized stages, remains mostly unpulled. Screening is the primary mechanism for that shift. No drug changes where a cancer is found. Screening does.

Adenocarcinoma, the subtype in Pope Leo's diagnosis, is notable in this context. It is now the most common form of non-small cell lung cancer, has a higher representation among never-smokers than other subtypes, and tends to present peripherally, where low-dose CT is effective. Stage at diagnosis still determines outcomes for adenocarcinoma as much as for any other subtype. Early-stage adenocarcinoma is among the most treatable cancers in the SEER database. Late-stage adenocarcinoma is not.

5. Who the system keeps missing

The Kentucky BRFSS analysis is valuable because it disaggregates the national average into its components, and the components tell a very different story by insurance status, geography, and race.

Screening uptake by insurance type in 2024: adults with military-related coverage screened at 39.05%, the highest of any group. Adults with employer-sponsored insurance screened at roughly 27%. Adults on Medicaid or Medicare screened at roughly 22 to 26%. Uninsured adults screened at 6.07%, and showed essentially no improvement between 2022 and 2024.

Screening by state ranged from 13.4% in South Dakota to 38.4% in Massachusetts. The highest-rate states clustered in the Northeast. Southern states, where lung cancer burden is high, consistently screened lower. A prior analysis of 2022 BRFSS data linked higher state screening rates to Medicaid expansion and greater radiology capacity. States that expanded Medicaid early showed systematically better screening penetration.

Screening by race: Hispanic, Asian, and Black adults all screened at lower rates than white adults and showed no statistically significant gains between 2022 and 2024.

Screening by age: uptake was 11.3% among adults ages 50 to 54, rising to just over 31% among those ages 75 to 79. The youngest eligible patients, who have the most life-years to gain from early diagnosis, are the least likely to be screened. The reasons are structural: primary care visits are less frequent among younger working-age adults, smoking history documentation is less complete, and awareness of eligibility is lower.

Insurance coverage design adds another layer. Medicare covers annual low-dose CT for beneficiaries ages 50 to 77 with at least a 20 pack-year history who currently smoke or quit within the past 15 years. The ACA requires most private plans to cover the initial scan without cost sharing. The problem is that the pathway does not end at the scan. Follow-up imaging, diagnostic bronchoscopy, tissue biopsy, and specialist consultations can carry copays, coinsurance, and deductibles under plans that cover only the initial screen. A patient who gets the scan and then receives a Lung-RADS 3 or 4 result may face out-of-pocket costs they cannot manage at exactly the moment they most need to engage.

New York addressed this directly. On December 4, 2025, Governor Hochul signed legislation requiring state-regulated insurers, including Medicaid, to cover follow-up screening and diagnostic services for lung cancer without cost sharing. The law takes effect January 1, 2027. It is one of the most specific policy interventions targeting the follow-up dropout point. Whether other states follow will depend on whether advocates can make the dropout data visible to legislators the way the ACS model made the death gap visible to researchers.

The never-smoker question is also real and growing. A 2026 review in the ASCO Educational Book estimated that 15% to 25% of lung cancers worldwide occur in people who never smoked. In the TALENT study, a prospective trial of never-smokers, the cancer detection rate was 2.6%, compared with 1.1% in the National Lung Screening Trial, and 96.5% of detected cancers were stage 0 or I. Several researchers have argued that excluding never-smokers cannot be justified given these detection rates. Others, including cancer center leadership at the University of Kentucky, have urged caution before widening criteria ahead of randomized mortality data. Both positions are defensible. Neither changes the near-term math: the largest available gain is among people who currently qualify and are not being reached.

6. What a working pathway looks like: the CommUnityCare model

The CommUnityCare Health Centers study is the most useful published evidence for safety-net operators trying to build a lung screening program.

CommUnityCare is a large FQHC system serving Central Texas. Between 2020 and 2023, its team built and ran a structured lung cancer screening implementation. Researchers published the results in the American Journal of Preventive Medicine in September 2024. The patient population was about 36% Latino and 18% African American. About 14% were uninsured, 48% were in a county medical assistance program, 27% were on Medicare or Medicaid, and 77% currently smoked.

The implementation used a multi-layer identification approach: EHR queries to surface likely eligible patients, direct provider referrals at the point of care, and outreach mailings to patients not recently seen. Identified patients were routed to a bilingual social worker who confirmed eligibility, conducted telehealth shared decision-making visits, coordinated CT scan scheduling, and offered smoking cessation support. For uninsured patients, financial assistance covered the cost of the scan. A board-certified chest radiologist reviewed all images for quality and refined initial reads, reducing false-positive rates. Patient navigators tracked Lung-RADS 3 and 4 findings through to follow-up workup.

The results:

Of 650 patients who completed shared decision-making, 636, or 98%, agreed to screening. Of those, 528 completed the CT, about 83%. Median time from referral to CT completion was 27 days. Completion exceeded 80% across subgroups, and differences by race, insurance status, and smoking status were not statistically significant.

Among screened patients, 88% had Lung-RADS 1 or 2 results (no significant finding). About 9% had Lung-RADS 3 (probably benign, short-term follow-up imaging recommended). About 3% had Lung-RADS 4 (suspicious, further diagnostic evaluation warranted). Seven patients, 1.3%, were diagnosed with cancer.

The referral pathway made a substantial difference in completion. Of 629 patients referred directly by providers, 579, or 92%, completed shared decision-making. Of 479 patients reached by mailed outreach, 108 were eligible and 71 completed shared decision-making, about 66%.

What CommUnityCare demonstrates is not that FQHCs have magical access to hard-to-reach populations. It is that when the pathway is engineered end to end, when eligibility is confirmed systematically, when the decision-making visit happens proactively, when the scan is financially covered, and when follow-up is tracked by a named person, a population that is mostly Latino, often uninsured, and predominantly current smokers will complete screening at rates the national surveys have not come close to approaching.

The lesson for health system operators is operational, not clinical. The clinical case for screening is settled. What is not settled, at most institutions, is who owns each step of the pathway.

What this means for health systems, FQHCs, and radiology teams

Pope Leo's diagnosis gave health systems a brief window of public attention. Here is what to do with it, in order of immediate impact.

**Measure what you have.** Ask your analytics team how many patients in your panel meet USPSTF criteria and how many had a documented shared decision-making visit and a low-dose CT in the last 12 months. If the answer is "we do not track this," that is your first finding. You cannot improve a number you cannot see.

**Build the referral pathway before the awareness campaign.** Awareness without capacity is waste. If a patient who calls after reading about Pope Leo's diagnosis cannot get an eligibility check, a shared decision-making visit, and a scheduled scan within 30 days, the attention converts to nothing. CommUnityCare achieved a 27-day median. That is a reasonable target.

**Assign a named person to Lung-RADS 3 and 4 follow-up.** This is where most programs fail. An abnormal scan with no follow-up owner does not save anyone. The navigator function does not need to be a high-cost clinician. It needs to be a person with a list who makes calls.

**Close the follow-up cost gap.** If your state has not moved on follow-up cost sharing, look at what New York did and make the financial case to your payer partners. The uninsured screened at 6.07% nationally. If your FQHC serves uninsured adults, the scan itself needs to be covered and the follow-up needs to be affordable, or the program is a funnel with a hole in it.

**Use the 47,000 as a planning figure.** It is a model, not a registry count. But it tells you the order of magnitude of what is not being done, and it gives benefits directors and payers a frame for the cost of late diagnosis versus the cost of an annual low-dose CT. The math is not close.

For radiology groups: the question is whether you have the capacity to absorb increased referrals, standardized Lung-RADS reporting, and a chest-trained reader on every low-dose CT. Volume will eventually arrive. The practices with a structured screening program in place will capture it.

For FQHCs: the CommUnityCare model is a proof of concept, not a template that transfers wholesale to every system. But its core elements, systematic EHR-based eligibility identification, telehealth decision-making, financial coverage for the scan, and navigator-managed follow-up, are replicable. The question is whether lung screening is a program with a budget and a workflow, or an aspiration waiting for someone to own it.

About

Jonathan Govette is Co-Founder and CEO of Oatmeal Health, which builds AI lung cancer diagnostics to catch cancers earlier in underserved communities. Oatmeal Health's LungAI is investigational and not FDA cleared. Jonathan writes The Oatmeal Bite for health system leaders, FQHCs, and radiology teams navigating the intersection of technology, policy, and population health.

Key References

• Vatican announcements of surgery and diagnosis, October 13 to 14, 2026. Surgery performed at Gemelli Hospital, Rome, by Dr. Stefano Margaritora. Adenocarcinoma confirmed on intraoperative examination. Stage not publicly disclosed. Reported by CBS News, NPR, and MedHeadlines.

• Bandi P, et al. Lung cancer deaths prevented and life-years gained from lung cancer screening. JAMA research letter, November 19, 2025, DOI 10.1001/jama.2025.19798. Used 2024 NHIS; estimated 12.76 million eligible adults; 18.7% up to date; modeled 62,110 deaths preventable at full uptake vs 14,970 at current uptake. Via The ASCO Post and AuntMinnie. Managed Healthcare Executive October 15, 2026, for mammography comparison (65% vs 18%).

• Burus T, et al. Lung cancer screening uptake using 2024 BRFSS. JAMA Internal Medicine research letter, April 27, 2026. 24.5% uptake nationally; state range 13.4% (South Dakota) to 38.4% (Massachusetts); uninsured 6.07%; military-related coverage 39.05%. Via AuntMinnie (April 30, 2026) and Managed Healthcare Executive (May 21, 2026).

• Rustagi A, et al. All-cause mortality in Veterans Health Administration lung cancer screening participants vs NLST participants. JAMA Network Open research letter, April 27, 2026. Via AuntMinnie.

• National Cancer Institute SEER Cancer Stat Facts: Lung and Bronchus Cancer, 2026 estimates using 2016 to 2022 data. Five-year relative survival: localized 65.5%, regional 38.2%, distant 10.5%. Stage distribution: localized 24%, regional 21%, distant 51%.

• American Cancer Society. Cancer Statistics 2026 press release. Estimated 229,410 new U.S. lung cancer cases and 124,990 deaths in 2026; approximately 20% of all cancer deaths. Overall five-year survival 28%, up from 15% in mid-1990s.

• CMS National Coverage Determination 210.14. Lung Cancer Screening with Low Dose Computed Tomography, effective February 10, 2022. Age 50 to 77, 20 pack-years, currently smoking or quit within 15 years, shared decision-making visit required before first scan.

• New York Governor's Office. December 4, 2025, signing of S2000A/A1195A. Requires state-regulated insurers including Medicaid to cover lung cancer follow-up screening and diagnostic services without cost sharing. Effective January 1, 2027.

• CommUnityCare Health Centers implementation study. Villalobos-Gonzalez C, et al. American Journal of Preventive Medicine, September 2024. 650 shared decision-making completers; 636 (98%) agreed to screen; 528 (83%) completed CT; median 27 days; 7 cancers detected (1.3%). Via Radiology Business (October 8, 2024) and ASCO Post (October 2024).

• Gorria T, et al. Never-smoker lung cancer screening review. ASCO Educational Book, 2026. TALENT study: 2.6% cancer detection rate; 96.5% stage 0 or I. Via AJMC.

• Jiang B, et al. AI-first reader workflow for lung cancer screening. European Radiology, March 2026. 1,252 UKLS scans; 1.9% of cases were isolated typical perifissural nodules requiring radiologist review; 0 of 57 typical perifissural nodules malignant. One author employed by Coreline Soft.

On September 1, 2026, the first day of Texas's new state fiscal year, three Medicaid directed payment programs went dark. Federal approval had not arrived. The money that Texas hospitals, physician groups, rural clinics, nursing facilities, and behavioral health providers had budgeted was not there.

The Texas Hospital Association estimated the daily cost at $27 million. For about two and a half weeks, that was the cadence: $27 million a day not flowing, and hospitals running their operations against a gap they had not planned for.

On September 17, Governor Greg Abbott announced that CMS had agreed to release nearly $12 billion. John Hawkins, president and CEO of the Texas Hospital Association, said the Medicaid crisis hanging over the state's hospitals had been averted. Abbott called the funding a major victory for Texas patients.

Both statements are accurate. Neither is the whole story. The September 17 announcement resolved an acute standoff. It did not resolve the structural dispute o…

Read more

The No Surprises Act took effect on January 1, 2022 with a clear mandate: stop patients from being blindsided by bills from out-of-network providers at in-network facilities. It largely worked. Patients are no longer the ones caught between a hospital and an insurer arguing over who pays.

What nobody fully anticipated was the machine that would grow up around the dispute resolution system meant to settle those arguments. Four years later, that machine has processed 2.5 million disputes in a single year, awarded providers roughly $15 billion in a single year, and generated $22.4 billion in total system costs since 2022. On October 8, 2026, Rep. Frank Pallone Jr. of New Jersey, one of the law's original authors, introduced the Lower Premiums, Faster Payments Act, which would replace the arbitration system with a price benchmark tied to 2019 contracted rates.

This article covers what the bill does, what the data says, who is fighting it, what is already changing, and what operators need to…

Read more

Lung cancer kills roughly 125,000 Americans a year. It causes more cancer deaths than breast, prostate, and colon cancer combined. A low-dose CT scan can find it early, when it is still treatable. That scan is covered, endorsed by guidelines, and available at hundreds of facilities across the country.

And yet a new analysis of the most recent Health Information National Trends Survey, published in the American Journal of Surgery, finds that only 15% of adults who qualify for lung cancer screening have discussed it with a clinician.

Not 15% who have been screened. 15% who have even had the conversation.

That number is the real bottleneck. Not the scan. Not the coverage. The talk.

What the Number Says

Researchers at the University of Cincinnati Cancer Center, led by senior author Robert Van Haren, MD, used HINTS, the National Cancer Institute's nationally representative survey of U.S. adults, to ask a direct question: had eligible adults spoken with a healthcare professional about lung cancer screening in the past year?

Out of roughly 1,500 adults who met USPSTF eligibility criteria, about 221 said yes.

That is 14.7%. The study rounds to 15%. Either way, it is approximately 1 in 7.

The USPSTF criteria for lung cancer screening are specific: adults aged 50 to 80, with at least a 20 pack-year smoking history, who currently smoke or quit within the past 15 years. This is not a broad population. These are people whose risk is documented, whose eligibility is defined, and whose insurer is generally required to cover the scan.

Among those eligible adults, the groups most likely to have had the conversation were current smokers and people already diagnosed with lung disease. That pattern makes sense. But it also means that the people who have already quit, who may assume their risk has faded, are the most likely to be skipped.

The American Cancer Society, in its most recent national dataset published in JAMA in November 2025, found that 18.7% of eligible adults were up to date with lung cancer screening as of 2024. BRFSS self-report data, which uses a slightly different eligibility window, puts the figure at 24.49%. Both numbers are improving. Twenty-one states showed significant increases since 2022.

But both numbers sit far below where they should be, and both are downstream of the counseling conversation. If only 15% of eligible adults have talked to a clinician about screening, the screening rate cannot meaningfully exceed that ceiling for long. At some point, the conversation precedes the scan.

The Comparison That Matters

Colonoscopy screening rates in the U.S. run between 70% and 80% for eligible adults. Mammography rates are similar. These numbers are not perfect, and access gaps remain for both. But the counseling conversation for colorectal and breast cancer screening has been normalized over decades of physician training, patient education, and public messaging.

Lung cancer screening was added to the USPSTF recommended list in 2013 and updated in 2021, when eligibility was expanded to include adults aged 50 to 80 and reduced the pack-year threshold from 30 to 20. Those changes nearly doubled the eligible population.

But awareness did not double. Training did not double. The counseling conversation did not follow.

The Centers for Medicare and Medicaid Services took an unusual step in 2015 when it required beneficiaries to complete a shared decision-making counseling visit before a covered lung cancer screening CT scan. This is the only such mandate for a preventive screening test in Medicare. The rationale: lung cancer screening carries real tradeoffs. False positives, unnecessary biopsies, incidental findings, and the psychological weight of a lung nodule finding all need to be discussed. The conversation is not a formality. It is part of the care.

The problem is that requiring a conversation before the scan assumes the conversation is happening. The Viana study shows it largely is not.

The Stage You're Catching

The reason screening matters comes down to what stage the cancer is at when you find it.

For non-small cell lung cancer, which accounts for roughly 80% to 85% of cases, the five-year relative survival rate when the disease is localized is 67%. That means 2 out of 3 people diagnosed at an early stage are still alive five years later.

When the disease has spread to distant organs, that number falls to 12%.

For small cell lung cancer, the gap is even wider: 34% five-year survival for localized disease, 4% for distant.

These numbers come from SEER data, compiled by the National Cancer Institute from 2015 to 2021. They represent averages across a large, heterogeneous patient population, and individual outcomes vary based on histology, mutation status, performance, and treatment access. But the direction is clear and consistent: early detection saves lives.

The American Cancer Society estimates that at 100% lung cancer screening uptake among eligible adults, approximately 62,110 deaths could be prevented over five years, with 872,270 life-years gained. At current uptake rates, the realized benefit is roughly a quarter of that: about 14,970 deaths prevented and 190,030 life-years gained.

That is the math of the counseling gap. Every eligible adult who never has the conversation is a person who may never get the scan, who may be diagnosed at stage IV instead of stage I, and whose odds drop from 67% to 12%.

Where the Conversation Breaks Down

Primary care physicians face real constraints. A standard outpatient visit runs 15 to 20 minutes. Chronic disease management, medication review, preventive care, and the patient's own list of concerns all compete for that time. Adding a counseling conversation about lung cancer screening that covers eligibility, benefits, risks, false positives, follow-up protocols, and smoking cessation requires time that is not always available and is often not separately reimbursed.

CMS does reimburse the shared decision-making visit for Medicare beneficiaries, using HCPCS code G0296. But billing and coding complexity, clinician unawareness of the code, and the administrative burden of documentation mean that the reimbursement often goes unclaimed.

The training gap is also real. Lung cancer screening expanded significantly in 2021 when USPSTF broadened eligibility. Clinicians who trained before that change may still be operating on older criteria, either referring too few patients or referring patients who no longer qualify under the previous, narrower guidelines. The 2021 expansion added an estimated 6 to 7 million newly eligible adults. Many of their primary care physicians have not been updated on the new criteria in any systematic way.

Among clinicians who do discuss lung cancer screening, a meaningful share report that they are uncertain how to counsel patients on the tradeoffs. False positive rates for low-dose CT are not trivial, particularly in high-prevalence areas with endemic fungal infections like histoplasmosis or coccidioidomycosis, where benign nodules can mimic early malignancy. Ordering a scan without discussing what a positive result might require, including follow-up imaging, PET scans, or biopsy, sets the patient up for anxiety and the health system up for downstream costs that may not have been intended.

The counseling conversation is not optional under Medicare. It is required. But the data show it is also rare.

Who Gets Left Out Entirely

Lung cancer screening uptake already reflects the broader inequities in U.S. preventive care. The counseling gap compounds them.

Among adults who are eligible for lung cancer screening, BRFSS data from 2024 show that uninsured adults are screened at a rate of 6.07%. Adults with military-related insurance are screened at 39.05%. The ratio between those two numbers is more than 6 to 1. The difference is not biology. It is access.

Screening uptake also varies by state to a degree that cannot be explained by population differences alone. Massachusetts reports 38.36% of eligible adults up to date on lung cancer screening. South Dakota reports 13.43%. That is a nearly threefold difference between two U.S. states.

The American Cancer Society analysis found that about 28.1 million adults who have ever smoked but do not currently meet USPSTF eligibility would still benefit from screening, particularly those who quit more than 15 years ago. The American Cancer Society's own 2023 guidelines expanded recommendations to anyone with at least a 20 pack-year history regardless of quit date. If the counseling rate among those who do qualify is 15%, the rate among those who are newly eligible or just outside eligibility criteria is likely even lower.

Rural adults face additional structural barriers. Lung cancer screening programs are concentrated in academic medical centers and urban hospital systems. A patient in a nonmetropolitan county who qualifies for screening may have no nearby facility with an accredited program, no clinician who regularly orders low-dose CT, and no system-level outreach to identify them as eligible.

The BRFSS data show that after adjustment, nonmetropolitan uptake (23.80%) is not significantly different from metropolitan (24.69%). But that national average masks local variation. In rural counties with high smoking prevalence and limited primary care access, the true gap is much wider.

What Closing the Gap Requires

Improving the counseling rate for lung cancer screening is not primarily a technology problem. It is an infrastructure and workflow problem.

Proactive outreach is one lever. Electronic health record systems that flag eligible patients and prompt clinicians at the point of care have shown measurable increases in referral rates. Order sets and decision support tools that automate the eligibility calculation, removing the clinician's need to remember the pack-year threshold and age criteria, reduce the friction that lets eligible patients leave without a referral.

Nonphysician-led counseling is another. The CMS requirement specifies a clinician visit, but trained nurses, respiratory therapists, social workers, and health coaches can conduct the shared decision-making conversation when it is structured around a validated decision aid and documented appropriately. Delegating the conversation to a trained support team member extends the reach without requiring a physician appointment for every eligible adult.

Telehealth-delivered counseling has shown promise in reaching patients who are unlikely to schedule an in-person visit. A 15-minute video visit to discuss eligibility, answer questions about the scan, and place the referral can reach patients who would not otherwise have the conversation.

Community health outreach, particularly through federally qualified health centers, rural health clinics, and community health workers embedded in high-risk communities, addresses the geographic and trust barriers that national averages obscure. These are the organizations with existing relationships in the communities where lung cancer rates are highest and screening rates are lowest.

The American Lung Association's "State of Lung Cancer" report and the ACS data make the same point from different angles: the infrastructure to find and screen people exists. The gap is in connecting eligible patients to the first conversation.

What This Means for Oatmeal Health

Oatmeal Health exists because lung cancer is most survivable when it is caught early, and most lung cancers are not caught early. The scan is one part of the solution. Getting eligible patients to the scan is another. But neither happens without the conversation that comes first.

The Viana study puts a specific number on the conversation gap: 15%. That number is not a rounding error. It represents roughly 10.8 million people who meet USPSTF criteria for lung cancer screening, qualify for covered scans, and have not had the talk with a clinician.

Our work is focused on closing that gap. That means identifying eligible patients, connecting them with the counseling conversation, and making it possible for health systems to deliver proactive lung cancer screening at scale, not as a reaction to a patient who asks, but as a systematic process that finds people before they find cancer.

The bottleneck is not the scan. The bottleneck is the conversation. Solving for that is what early detection in lung cancer actually looks like.

Jonathan Govette is Co-Founder and CEO of Oatmeal Health, a company focused on expanding access to lung cancer screening through proactive patient identification and health system partnership. Oatmeal Health works with health centers, health systems, and payers to close the gap between screening eligibility and actual screening completion.

Key References

• Viana SW, et al. "National analysis of factors associated with patient-clinician discussions about lung cancer screening." *The American Journal of Surgery*. 2026. Senior author: Robert Van Haren, MD, University of Cincinnati Cancer Center. (HINTS data; n approx. 1,500 eligible adults; 15% discussion rate)

• American Cancer Society. "Latest ACS Lung Cancer Data: Only 1 in 5 Eligible Adults in U.S. Screened for Lung Cancer." JAMA, November 19, 2025. Pressroom: pressroom.cancer.org/2025-lung-cancer-data

• National Cancer Institute. BRFSS lung cancer screening uptake analysis. PMC13122494. 2024 data; n=26,104 respondents.

• American Cancer Society. Lung cancer survival rates by stage. SEER data 2015-2021 (NSCLC) and 2012-2018 (SCLC). cancer.org. Accessed June 2025.

• U.S. Preventive Services Task Force. "Recommendation: Lung Cancer: Screening." Updated 2021. uspreventiveservicestaskforce.org

• Centers for Medicare and Medicaid Services. HCPCS code G0296: Counseling visit for shared decision making on lung cancer screening. CMS, 2015.

• AuntMinnie. "How useful is shared decision-making for lung cancer screening?" 2024. auntminnie.com/clinical-news/ct/article/15627943/

• Oatmeal Health. "Lung Screening: Only 1 in 6 Get the Talk." October 11, 2026. oatmealhealth.com/lung-screening-only-1-in-6-get-the-talk/

A Note on the Numbers: Why Survey Estimates Diverge

One detail worth addressing directly: the three most-cited figures for lung cancer screening uptake in the U.S. do not match.

The Viana HINTS study finds 15% of eligible adults have discussed screening with a clinician.

The ACS/JAMA analysis finds 18.7% of eligible adults were up to date on screening as of 2024.

The BRFSS analysis puts self-reported up-to-date screening at 24.49%.

These numbers measure different things and use different instruments. The HINTS counseling figure measures whether a conversation happened, not whether a scan occurred. The ACS figure uses NHIS (National Health Interview Survey) data. The BRFSS figure is based on a state telephone survey with a different age range and different eligibility proxies.

None of them are wrong. They are looking at different slices of the same problem.

What they share: every estimate falls well below the rates at which other cancer screenings are completed. And every estimate puts the counseling conversation rate, the precondition for informed, shared, Medicare-compliant screening, lower than the screening rate itself.

That last point is worth sitting with. Some eligible adults are getting screened without the full shared decision-making conversation the guidelines recommend and CMS requires for Medicare beneficiaries. They may have had a physician recommend a scan without the structured discussion. They may have self-referred. They may be Veterans with access to VA programs that structure outreach differently.

But for the majority of the 85% of eligible adults who are not up to date on screening, the bottleneck is not refusal. It is never being asked.

The American Cancer Society's analysis finds that at current uptake rates, about 14,970 deaths are prevented over five years. At 100% uptake, that number rises to 62,110. The difference between those two figures is 47,140 lives over five years. That is roughly 9,400 people a year who could be saved if the eligible population were actually reached and screened.

Getting to 100% uptake is not a realistic near-term target. But moving from 18.7% to 40% is not unrealistic, and even partial progress at scale has significant mortality impact. The question is which intervention, at which point in the care pathway, moves the needle fastest.

The evidence points toward the conversation.

On October 1, 2026, Tiny Health closed a $33 million Series B. The round was oversubscribed. B Capital led it. Stanford Medicine, Mayo Clinic, and Massachusetts General Hospital are on the enterprise partner list. More than 6,000 practitioners use the platform.

And the CEO describes the product as a wellness test.

That gap between the institutional partnerships and the regulatory label is not a contradiction. It is a strategy. It is also a lens for understanding one of the more important questions in diagnostics right now: when does clinical adoption require clinical evidence? The answer is not as obvious as most procurement policies assume.

This piece walks through the Tiny Health round, the evidence behind the test, the regulatory lane the company chose deliberately, and the oncology science that shows where the microbiome field is actually heading. It ends with a practical checklist for the administrators, medical directors, and benefits leaders who will be pitched something like this next quarter.

1. The round in plain numbers

Tiny Health raised $33 million in a Series B led by B Capital. Existing investors Spero Ventures, The Venture City, and Overwater Ventures participated. New capital came from Black Opal Ventures, Denver Ventures, Pave Health Ventures, Alumni Ventures, Gaingels, and Pari Passu Ventures. B Capital's Nick Whitehead, a senior principal who called the gut microbiome "one of the most significant and under-explored frontiers in human health," joins the board.

The round brings total funding to $46 million. Working backward: a $4.5 million seed in 2021 and an $8.5 million Series A in late 2023 account for $13 million. The Series B is roughly 72% of everything the company has ever raised. Three-quarters of its capital arrived in one round.

The stated use of proceeds covers three areas. First, scaling TinyAI, an analytical engine trained on approximately 200,000 longitudinal gut and vaginal microbiome profiles. Second, launching a $5 million Microbiome Research Program that invites academic and clinical researchers to use Tiny tests and data. Third, expanding research into four clinical areas: inflammatory bowel disease, perimenopause and menopause, oncologic drug response, and GLP-1 receptor agonist interactions.

Notice the shape of that list. One item is a product. Two are research. A company that calls itself a wellness test is spending a meaningful share of new capital on the work a diagnostics company does when it wants to cross the payer line.

The company did not disclose a valuation or an absolute revenue figure. CEO Cheryl Sew Hoy told MedCity News that revenue has more than doubled each year. The company has roughly 40 employees and plans to add 30, a 75% headcount expansion funded by this round. The direct-to-consumer tests cost $249 each. The Tiny+ subscription costs $399 per year for two tests; more than half of customers subscribe. Adult test volume now exceeds pediatric test volume. Half of revenue comes from DTC.

These are directional figures from a private company. Treat them as such.

2. The 54-infant trial that anchors the narrative

The centerpiece of Tiny Health's clinical story is a randomized controlled trial published in Pediatric Allergy and Immunology in 2025. The citation is Nieto et al., doi: 10.1111/pai.70182. The paper's title is "Improving immune-related health outcomes post-cesarean birth with a gut microbiome-based program: A randomized controlled trial."

The result is a striking one. Infants in the intervention arm had 83% lower odds of developing eczema compared to controls.

What the trial actually shows deserves a careful read.

The study enrolled C-section infants and divided them into an intervention arm and a control group. The intervention combined at-home microbiome testing, strain-specific probiotic recommendations, and parental education. In the intervention arm, 64% of infants received Bifidobacterium-containing probiotics. In the control arm, 3% did. The follow-up window is described as short, and the paper itself notes the small sample size as a limitation. The publicly available study page does not state the odds ratio numerically, does not give a confidence interval, and does not report a p-value.

A modeled scenario circulating alongside the trial projects 260 fewer eczema cases per 1,000 pediatric patients. That number comes from Dr. Elisa Song, a collaborator and Tiny Health's chief medical officer, speaking with NutraIngredients. The article identifies it explicitly as a model, not a trial result. Both figures show up in vendor materials. Buyers should know which is which before a purchasing decision.

None of this makes the 83% result false. A small randomized trial with a strong effect is a legitimate reason to run a larger one. Tiny Health says a second RCT is active. The appropriate framing for the existing evidence is: early, company-run, and promising. The follow-up trial will tell a more complete story.

What the existing evidence does not support: use in clinical practice as a validated diagnostic, payer coverage, or comparison to FDA-cleared devices. The 2025 Lancet Gastroenterology and Hepatology international consensus statement on the microbiome, which surveyed the global literature, concluded that current gut microbiome data may be useful for research purposes but that "there is still insufficient evidence to apply them in clinical practice." The same panel advised against reports recommending therapies, including dietary interventions, because those recommendations fall outside evidence-based practice.

That is the peer-reviewed consensus view, as of 2025. The Tiny Health trial is a meaningful contribution to a field where that consensus may shift. But as of today, the field's own experts say it has not shifted yet.

3. "We are a wellness test": a regulatory lane, chosen deliberately

Sew Hoy's phrase is not a slip. Every word is intentional. A product positioned as wellness does not claim to diagnose, treat, or prevent disease. That positioning shapes everything downstream: who regulates it, who pays for it, and what evidence a buyer can reasonably demand before putting it near a patient.

The regulatory backdrop changed materially in 2025. On March 31, 2025, the U.S. District Court for the Eastern District of Texas vacated the FDA's final rule on laboratory-developed tests (LDTs). The consolidated cases were American Clinical Laboratory Association v. FDA and Association for Molecular Pathology v. FDA. The court held that LDTs are services, not devices under the Food, Drug and Cosmetic Act, and that FDA's device authority does not extend to professional testing services offered from a single laboratory. CLIA, administered by CMS, remains the oversight framework.

FDA chose not to appeal and has issued a notice of intent to rescind the rule. The ruling is narrow: it applies to tests developed, validated, and offered within a single laboratory. Tests distributed across multiple laboratories, or tests built on FDA-regulated components, face unresolved questions. But for a test that generates results from a single CLIA-certified lab and issues wellness-framed reports, the ruling means no FDA submission is required.

That is not a critique of Tiny Health specifically. It describes the lane. Roughly 31 DTC microbiome testing companies worldwide operated as of a December 2023 University of Maryland report. About half of them also sell supplements and recommend them in test reports. Twelve of the 27 US-based companies in that analysis claimed CLIA-compliant labs. None of those products carry FDA clearance, because FDA clearance is not required in this lane.

A 2026 study published in Communications Biology tested seven DTC gut microbiome services using the same standardized fecal sample, prepared according to NIST reference methods. The study found major discrepancies in species identification and relative abundance between providers. Critically, the variability between providers was approximately the same magnitude as the biological variability between different human donors. The authors' conclusion: "Analytical performance is a prerequisite for making sound clinical recommendations."

A wellness label answers to a marketing standard and a laboratory's own quality system. A cleared device answers to FDA on intended use and clinical validity. A covered service answers to payers on outcomes. A product can clear one bar without clearing the others. When an enterprise partner list reads Stanford, Mayo, and Mass General alongside a $249 DTC consumer kit, the institutional affiliations do not automatically validate the product for clinical use. They validate it for research and for adoption by practitioner communities that have made their own judgment calls.

4. Why the B2B arm matters more than the consumer kit

The DTC story is easier to tell, but the more consequential story is the practitioner channel.

Powered by Tiny, the B2B platform, grew fourfold in the past year. More than 6,000 practitioners use it. Forty-seven enterprise partners are named, including Stanford Medicine, Fullscript, Mayo Clinic's Executive Health and Longevity Programs, Atria Institute, and Massachusetts General Hospital. Nick Whitehead of B Capital told MedCity News that when the firm surveyed practitioners who had evaluated multiple microbiome tests, Tiny Health consistently rated highest.

That is a real distribution advantage. It also illustrates how clinical adoption usually works in practice, and the pattern matters for anyone responsible for a purchasing decision.

A diagnostic tool rarely enters a workflow because a randomized trial told a committee to buy it. It enters because clinicians in a specific niche, here executive health, longevity medicine, and integrative practice, adopt it first. Peers follow. Evidence arrives later, sometimes much later. Imaging AI followed an almost identical path. Many radiology AI tools reached daily use in US health systems before any head-to-head prospective evaluation was published. The FDA cleared more than 1,600 AI and machine learning devices through 2025. The majority of those cleared devices have no published post-market performance data.

6,000 practitioners is a distribution base, not a clinical outcome. Adoption answers a different question than evidence does. The adoption number tells you the product is usable and that a defined community of clinicians has found value in it. It does not tell you whether the product changes patient outcomes, lowers downstream cost, or reduces harm at a population level. Those are separate claims, and they require separate evidence.

When a vendor presents adoption and evidence side by side in a pitch deck, ask which question each data point answers. The first is a sales metric. The second is a clinical metric. They are not interchangeable, and a health system that conflates them is making its procurement decisions on the wrong basis.

5. The oncology signal: the science is moving faster than the products

None of this is a case that the microbiome does not matter clinically. The case that it does matter, in specific, measurable contexts, is getting stronger.

The strongest recent signal is in oncology. In January 2026, Nature Medicine published the full results of the FMT-LUMINate trial. The phase 2, multicenter, open-label study was run at five academic centers across Canada. Lead investigator Dr. Arielle Elkrief of Universite de Montreal and colleagues combined fecal microbiota transplantation with immune checkpoint inhibition in two parallel cohorts.

The trial enrolled 40 treatment-naive patients: 20 with non-small cell lung cancer and PD-L1 expression of 50% or higher, and 20 with cutaneous melanoma. The protocol began with bowel preparation followed by 36 to 40 oral FMT capsules from healthy donors. Within one week, NSCLC patients started pembrolizumab monotherapy; melanoma patients started nivolumab plus ipilimumab.

The results in the NSCLC cohort were striking. Objective response rate: 80%. Disease control rate: 95%. One-year overall survival: 100%. There were no grade 3 or higher adverse events in the NSCLC group.

The melanoma results were more complex. Objective response rate: 75%. One-year overall survival: 79%. But 65% of melanoma patients experienced severe toxicities. A 15% incidence of myocarditis, linked to recipients of Prevotella-rich donors, was reported.

The authors are explicit about the limitations. It was a small, single-arm study. Multiple donors were used, introducing variability. Enrollment was affected by pandemic-related adjustments. There is no control arm, so the response rates cannot be compared directly with standard immunotherapy alone. They call for larger mechanistic studies.

An 80% response rate in 20 NSCLC patients is a hypothesis-generating result. It is not a standard of care, and it is not ready for clinical adoption. But it is a real signal, generated by credible investigators at academic centers, published in a top-tier journal, with a defined mechanism (donor microbiome composition modulating immunotherapy response) and a defined population (PD-L1 high NSCLC). That is exactly the kind of early evidence that earns a phase 3 trial.

Tiny Health's stated research agenda includes oncologic drug response. The company has 200,000 microbiome samples. If that data produces controlled evidence linking microbiome profiles to immunotherapy outcomes, the company will have earned a different regulatory lane. Until then, the oncology literature and a wellness kit share a field without sharing an evidence base.

6. What every buyer should ask before a microbiome tool enters a clinical workflow

Procurement leaders, medical directors, and benefits administrators get pitched tools like this frequently. The questions below are not specific to Tiny Health. They apply equally to imaging AI, to liquid biopsy tests, to any new diagnostic tool looking for a home in a clinical workflow.

First: What is the intended use, in the vendor's own regulatory language?

"Wellness" and "diagnostic" are not synonyms. The intended use determines which regulatory lane the product is in, what evidence the company is required to produce, and what a health system can be held accountable for when it recommends the product to a patient. The contract should match the claim. If the vendor says "wellness" but the sales pitch implies clinical utility, that gap is a liability.

Second: What is the evidence, by design?

Ask for the full citation for every study named. Ask specifically for the odds ratio, the confidence interval, the sample size, the follow-up duration, the enrollment criteria, and the funder. A company-funded trial is not automatically bad evidence. It is a reason to look for independent replication. If no independent replication exists, ask why and when one is planned.

Third: Who pays, and why?

Tiny Health's tests are not covered by insurance. A $249 cash-pay kit reaches a different population than a covered clinical service. If the evidence for a product comes from a specific population (infants born via C-section to health-conscious parents who seek out a $249 kit), applying that evidence to a general clinical population requires explicit justification.

Fourth: What happens to the data?

Tiny Health holds approximately 200,000 microbiome samples and is building a predictive model on them. The $5 million research program will give academic researchers access to tests and data. For any product that involves patient samples or patient-level data, the contract should specify consent procedures, secondary-use limitations, data retention terms, de-identification standards, and what happens to the data if the company is acquired or changes its business model.

Fifth: What is the pathway to reimbursement, and when?

CEO Sew Hoy has said publicly that reimbursement depends on continued scientific validation and clinical evidence. That is a credible answer. It is also a clear statement that the product is not in a reimbursed clinical workflow today. A tool that depends on out-of-pocket spend structures who benefits from it. Health systems that adopt cash-pay diagnostics for executive health programs are making a different choice than systems that want to close access gaps.

7. The lung cancer screening connection

At Oatmeal Health, we focus on a specific population: patients who face the highest risk of late-stage lung cancer diagnosis, and the systems that serve them. The FMT-LUMINate result matters directly to that work.

An 80% response rate in PD-L1-high NSCLC patients receiving FMT plus pembrolizumab is a meaningful signal in a disease where early detection is still the most powerful lever we have. Most lung cancers are still found late, when treatment options are limited and survival is poor. Improving treatment response in those late-stage patients matters enormously. If the microbiome is a real variable in immunotherapy response, then the populations we work with, who are already at elevated risk from detection delays, have the most to gain from getting this right.

The path from FMT-LUMINate to a clinical standard of care is long and requires controlled trials, donor standardization, and replication across populations. We are watching that path carefully, and we will report on it as the evidence develops.

What the Tiny Health round illustrates is the shape of the moment. Venture capital is moving into the microbiome field with real capital and real institutional partners. The science is producing genuine signals. The evidence gap between those signals and clinical-grade tools is real and measurable, but it is narrowing in some areas, particularly oncology. The question is whether the capital and the evidence are converging on the same problems.

The 54-infant eczema trial and the 20-patient NSCLC FMT trial are not the same kind of evidence. They are not asking the same question or reaching the same population. But they both point toward a field that is serious about turning observation into intervention. That is worth tracking.

About

Jonathan Govette is Co-Founder and CEO of Oatmeal Health, a company dedicated to making lung cancer screening accessible to every patient who needs it. He writes about healthcare diagnostics, payer policy, and the evidence gap between clinical promise and clinical practice. This newsletter comes out daily.

Key References

- Nieto P, et al. Improving immune-related health outcomes post-cesarean birth with a gut microbiome-based program: A randomized controlled trial. Pediatric Allergy and Immunology. 2025;36:e70182.
- Elkrief A, et al. Fecal microbiota transplantation plus immunotherapy in non-small cell lung cancer and melanoma: the phase 2 FMT-LUMINate trial. Nature Medicine. January 2026.
- NutraIngredients. "Tiny Health Raises $33M Series B to advance microbiome intelligence." October 1, 2026.
- MedCity News. "Tiny Health secures $33M to expand microbiome testing." September 2026.
- Lancet Gastroenterology and Hepatology. International consensus statement on gut microbiome applications in clinical practice. 2025.

Yet three other numbers barely moved. Predictive AI use in hospitals went from 66% to 71% across the same two survey years. Independent hospitals still trail system-affiliated hospitals by about 49 points. And the first major claims study of billable medical AI found that a handful of procedures drove nearly all volume, concentrated in wealthy, urban, academic markets. Use is rising faster than proof, and access is rising slower than use.

This is the October 2026 edition of a series built on the Wu et al. NEJM AI claims study of clinical AI adoption. It rebuilds the picture from 2023 to 2026 using newly checked sources: the FDA device list, AMA physician surveys, ASTP/ONC hospital data, Rock Health and Menlo Ventures funding data, and CMS and AMA reimbursement changes. It is written for health system leaders, FQHCs, radiologists, and benefits teams who have to decide what to buy, what to measure, and what to ignore.

1. The baseline: what the claims study found

The study that anchors this series, "Characterizing the Clinical Adoption of Medical AI Devices through U.S. Insurance Claims," came from Kevin Wu, Eric Wu, Brandon Theodorou, Weixin Liang, Christina Mack, Lucas Glass, Jimeng Sun, and James Zou at Stanford, the University of Illinois Urbana-Champaign, and IQVIA. NEJM AI published it on November 9, 2023, with funding from the Chan-Zuckerberg Initiative. It asked a simple question: which medical AI tools are clinicians actually billing for?

The authors tracked CPT codes created for medical AI in the IQVIA PharMetrics Plus for MedTech claims database, which covers more than 210 million enrollees, mostly in commercial plans. They analyzed 11 billion claims from January 1, 2018 to June 1, 2023. They found 16 medical AI procedures billable under 32 unique CPT codes. At the time, the paper notes, more than 500 AI devices had gone through FDA evaluation.

The pattern was concentration. Only 4 of the 16 procedures had more than 1,000 total claims, and only coronary artery disease analysis and diabetic retinopathy screening had passed 10,000. Coronary artery disease analysis (HeartFlow FFRCT, CPT 0501T to 0504T) led with 67,306 claims, followed by diabetic retinopathy (LumineticsCore, CPT 92229) at 15,097, Cleerly's coronary CT analysis at 4,459, and liver MR (Perspectum LiverMultiScan) at 2,428. Fifteen of the 16 procedures were created in 2021 or later, and the median procedure was about 374 days old.

📊 Data point: 15 of 16 billable medical AI procedures were less than three years old when the study window closed in mid-2023, and only 4 had more than 1,000 claims.

The geography was the part that should change how leaders buy. The authors looked at 22,704 zip codes with at least one institutional provider, and 2,182 had at least one medical AI billing. In a logistic regression, a zip code with an academic hospital was 17 times more likely to have AI billing, a metropolitan zip code 5.25 times more likely, and a high-income zip code (median household income above $100,000) 1.45 times more likely. All three were significant at P<0.001. Of zip codes with an academic hospital, 71% had at least one AI billing, versus 9% of those without.

📊 Data point: 71% of zip codes with an academic hospital had AI billing in the study, compared with 9% of zip codes without one.

Here is the full list. The 16 procedures and their 32 CPT codes, with claims through June 1, 2023, are below. Nearly everything sits in Category III, the AMA's temporary tier for emerging technology, which is why most had no Medicare price.

The second half of the list shows how thin the long tail was. Eight procedures had 237 claims or fewer, and two had none.

Pricing showed how payers valued these tools. CMS set the national payment for HeartFlow FFRCT at $930.34 in an office setting in November 2021, and Anthem's median negotiated rate in California and New York was $909.77. For diabetic retinopathy screening (CPT 92229), CMS paid $45.36 in 2021 while the median private rate was $127.81. The paper also noted that 92229 carried a practice expense relative value unit of 1.34 versus 0.53 for its non-AI counterpart, which shifts value from physician work toward the cost of the device.

The authors were explicit about limits. Their database captures a fraction of total claims (about 40% of U.S. claims by their estimate) and covers commercially insured patients under 65. It misses AI billed under non-AI-specific codes, AI paid through Medicare's inpatient NTAP program such as Viz.ai, unreimbursed pilots, and software bundled into hardware. They wrote that the claims counts should be read mainly as relative magnitudes over time.

One lung-related detail is worth keeping. CPT 0721T and 0722T, quantitative CT tissue characterization with Optellum's Virtual Nodule Clinic as the example product, had 4 claims in the dataset, effective July 1, 2022. In 2023, billable AI in lung nodule assessment existed on paper and almost nowhere in practice.

A later study using Medicare data confirms the pattern. Zink and colleagues, writing in The American Journal of Managed Care in May 2024, tracked CT-based fractional flow reserve (FFRCT), the first AI-enabled software Medicare paid for, using 100% of fee-for-service claims from 2018 through 2022. Annual claims grew from 1,083 to 12,363, and billing hospitals grew from 66 to 323. But the top 20% of hospitals submitted 67.5% of claims, and only about 17% of hospitals that performed coronary CT angiography also billed for FFRCT. Billing hospitals were larger (438 versus 241 beds) and more often part of a health system (86.9% versus 74.8%). The authors concluded that growth was concentrated among well-resourced hospitals and clinicians.

Every one of those findings is a snapshot from 2023 or earlier. The rest of this edition asks what changed.

2. Devices: from 221 authorizations a year to more than 1,600 on the list

Start with supply. The 2023 claims paper noted that more than 500 AI devices had been through FDA evaluation. In August 2024, MedTech Dive reported that the FDA had authorized 950 AI or machine learning-enabled devices, with more than three-quarters in radiology. The FDA's July 2025 list update added 211 devices cleared since its September 2024 update, about 81% of them in radiology, according to AuntMinnie. In August 2026, OncoDaily reported that the FDA list held more than 1,500 devices, and by September 2026 the FDA's own page says the agency has authorized over 1,600. The FDA page does not break the total down by year or specialty.

For annual counts, an independent tracker from Innolitics, summarized by PharmaDossier, puts 510(k) clearances at 221 in 2023, 253 in 2024, and 295 in 2025. That is a 33% increase from 2023 to 2025. In 2025, radiology accounted for 211 of the 295 clearances, or 71.5%. PharmaDossier also cites 1,451 cumulative devices through December 31, 2025, with 1,104 in radiology (76%).

📊 Data point: Annual AI clearances rose from 221 in 2023 to 295 in 2025, a 33% increase in two years, per an independent 510(k) tracker.

Two cautions apply. First, the annual series and the cumulative figure come from different trackers and do not sum to each other; the source itself notes this. Second, the FDA does not publish a per-year radiology series on its list page, so radiology share by year here is the tracker's, not the agency's. The direction is not in doubt. The precise counts are.

What matters for buyers is not the count. It is that radiology still holds roughly three-quarters of the list, which means the evidence base for AI outside imaging is thinner than the headline total suggests.

3. Physicians: 38% to 81% in three years

The fastest adoption curve is among clinicians. The AMA's annual physician survey found that 38% of physicians used AI in a professional context in 2023. In the survey released March 12, 2026, that figure was 81%, based on nearly 1,700 physicians surveyed in January and February. The average number of use cases per physician rose from 1.1 to 2.3.

The most common uses are documenting clinical care and summarizing medical research. Attitudes moved with usage. The share who believe AI gives physicians an advantage in caring for patients rose from 65% to 76%. The share who find more AI use more exciting than concerning rose from 30% to 37%, while 40% are equally excited and concerned.

📊 Data point: Physician AI use went from 38% (2023) to 81% (2026), and average use cases per physician from 1.1 to 2.3.

The concerns are specific. About 88% report some level of concern about skill loss, and 28% are personally concerned. Patient privacy is the only area where more physicians expect harm than help (41% versus 13%). And 85% want a say in how AI is adopted in their practice. AMA CEO John Whyte said AI "has quickly become part of everyday medical practice."

Read the 81% carefully. It counts any professional use, and the leading uses are documentation and research summaries, not diagnosis. That is real adoption, but it is not the same as adopting diagnostic AI that changes a clinical decision.

4. Hospitals: use is up, but the divide is persistent

Hospital data comes from ASTP/ONC, using American Hospital Association survey data. Among non-federal acute care hospitals, 66% reported using predictive AI integrated into their EHRs in 2023 and 71% in 2024. The 2024 numbers by hospital type show the divide: 86% of health system-affiliated hospitals, 81% of urban hospitals, 56% of rural hospitals, and 37% of independent hospitals.

The ASTP/ONC tables allow a direct comparison across the two years. Healthcare Brew, citing the same ASTP brief, reports that 96% of large hospitals used predictive AI in 2024 versus 59% of small hospitals. Rural use rose from 48% to 56%. Urban use rose from 77% to 81%. Multi-hospital system members went from 81% to 86%, and independent hospitals from 31% to 37%.

📊 Data point: Urban versus rural predictive AI use was 77% versus 48% in 2023 and 81% versus 56% in 2024. The gap narrowed from 29 points to 25 points.

This matters because the common story is that the divide is widening. On the 2023 to 2024 ASTP/ONC data, it narrowed slightly by urban and rural status (29 points to 25) and by system affiliation (50 points to 49). The divide is still large and ASTP/ONC calls it persistent. It is not, on this data, growing. A projection that it will widen needs a newer data release to support it. ASTP/ONC's Data Brief 80 (September 2025) is the latest hospital release found for this edition, and it covers 2023 and 2024 only.

A separate analysis by Oklahoma State University's Brian Whitacre, reported by KGOU on August 19, 2026 and still under peer review, uses 2023 AHA data from 1,301 urban and 1,018 rural hospitals and reports 51% rural versus 81% urban. The rural figure differs from ASTP/ONC's 48% because the samples and definitions differ. Whitacre's finding on purpose is the more useful one: rural hospitals are less likely to use AI for clinical decisions or treatment recommendations, equally likely to use it for billing and operations, and almost twice as likely to use models from their EHR vendor as to build their own. He argues for a "HITECH 2.0" program to help rural hospitals adopt and evaluate AI.

5. The governance gap: who checks the model?

Adoption is one question. Evaluation is another. ASTP/ONC also asked hospitals whether they test predictive models on their own data. In 2024, 82% of adopting hospitals reported checking accuracy, 74% checked for bias, and 79% conducted post-implementation monitoring.

The breakdown by hospital type is where the divide shows up again. Local accuracy evaluation among urban hospitals rose from 65% in 2023 to 75% in 2024, while independent hospitals went from 44% to 41%. Local bias evaluation rose from 47% to 61% at urban hospitals and from 30% to 31% at independent ones.

📊 Data point: Independent hospitals that check model accuracy on local data fell from 44% (2023) to 41% (2024), while urban hospitals rose from 65% to 75%.

For a health system leader, this is the sharpest finding in the dataset. The hospitals with the least capacity to adopt AI are also the least likely to test it. A model validated at a large academic center does not automatically perform the same on an independent hospital's patients. Under-resourced facilities inherit vendor claims they cannot check.

A JAMA Network Open study published December 12, 2025 by Jordan Everson, Paige Nong, and Chelsea Richwine adds the generative AI picture from the 2024 AHA IT Supplement. About 31.5% of hospitals were already using generative AI integrated with their EHR, and another 24.7% planned to within a year. System members were at 38.5% versus 16.3% for independent hospitals, urban hospitals at 35.5% versus 25.4% for rural, and major teaching hospitals at 53.9%. The authors also found that hospitals running all three local evaluation practices (accuracy, bias, and post-deployment checks) were slower to adopt than those running only one, and they warned of a digital divide in which some hospitals adopt without full safeguards.

The fastest-growing use cases also tell a story. In the same ASTP/ONC brief, using predictive AI to simplify or automate billing rose from 36% to 61% of adopting hospitals between 2023 and 2024, and scheduling from 51% to 67%. The money is moving to operations first.

6. The money: funding and spending, 2024 to 2025

Capital followed the same curve. Rock Health reports U.S. digital health venture funding of $10.9 billion in 2023 and $10.5 billion in 2024, then $14.2 billion in 2025, up 35% and the highest total since 2022. The first half of 2026 brought $7.4 billion across 244 deals, up from $6.4 billion in the first half of 2025 on a nearly flat deal count (245 then). Median deal size rose to $14 million, the highest since 2022. Deal count fell about 5%, from 509 to 482, so the dollars concentrated: average deal size rose from $20.7 million to $29.3 million. AI-enabled companies took 54% of 2025 funding, up from 37% in 2024. Rock Health no longer reports an AI-enabled share for 2026 because AI has become ubiquitous among digital health companies.

📊 Data point: Megadeals of $100 million or more took 22% of capital in 2024, 42% in 2025, and 45% in the first half of 2026. AI-enabled companies went from 37% of funding in 2024 to 54% in 2025.

Menlo Ventures' 2025 State of AI in Healthcare report, published October 21, 2025 and based on a Morning Consult survey of more than 700 executives in August and September 2025, puts 2025 healthcare AI spending at about $1.4 billion, nearly triple the prior year. No 2026 edition had been published when this was written. Of that, $600 million went to ambient clinical documentation and $450 million to coding and billing automation. Menlo reports that 27% of health systems, 18% of outpatient providers, and 14% of payers have deployed domain-specific AI.

Notice what that spending buys. Documentation and billing are the largest categories. Those are the same use cases that lead the physician and hospital data. Diagnostic AI, the category the FDA list is dominated by, is a smaller share of the dollars than it is of the device count.

7. Reimbursement: the Category I moment

The reimbursement story changed on January 1, 2026. CPT 75577, a new Category I code for AI quantification and characterization of coronary atherosclerotic plaque from coronary CT angiography, took effect. AuntMinnie reports that the 2026 Medicare physician fee schedule rate is more than $1,000 for physician offices and imaging centers, and the hospital outpatient rate is $950.50.

The code list changed in the same update. AMA's 2026 CPT set added 288 codes, deleted 84, and revised 46, and MedCentral's review of the AI-related changes names four: 75577, 0992T and 0993T (perivascular fat analysis for cardiac risk, without and with a concurrent heart CT), and 0972T (multispectral imaging for burn wounds). It also reports that 75577 replaced the four temporary codes 0623T to 0626T, the Cleerly codes in the 2023 study, which were deleted on January 1, 2026.

The 2026 and 2027 pipeline is bigger than four codes. AMA's Category III long descriptor document, updated October 2, 2026, lists 19 codes whose descriptors name augmentative or assistive algorithmic analysis or computer-aided analysis. Only two are effective now (July 1, 2026): 1036T and 1041T. The other 17 carry a January 1, 2027 effective date in the document's release table, although some of its narrative text also mentions October 1, 2026, so confirm the date in the official code set before billing. Cancer risk and prognosis is the largest group.

The cardiac, PET, and angiography codes cover tasks that already have FDA-authorized products behind them.

The remaining five span brain, breast ultrasound, intraoperative tumor imaging, and neurostimulation.

The same AMA source names seven earlier Category III codes for AI-enabled services that took effect in 2025: 0902T for electrocardiogram measurements, 0932T for echocardiogram measurement, 0877T to 0880T for medical chest imaging, and 0898T for image-guided prostate biopsy. AMA gives these as examples, not a complete list.

A caution on counts. Adding the 19 codes in AMA's current document to the 7 AMA names for 2025 gives at least 26 AI-related Category III codes beyond the 2023 study's list, and the 2023 study's 32 codes are not all still active (0623T to 0626T were deleted). That is a count of what these AMA sources name, not an official total, and no source found for this edition publishes one. Anyone quoting a single number should name the source and date. Looking ahead, a September 2026 analysis of the CPT Editorial Panel agenda by Medikode, citing Madhani Healthcare Consulting, counted 4 of 49 proposed Category I codes and 3 of 40 proposed Category III codes as involving AI or algorithmic analysis. Nothing from that meeting is in a live code set.

This is the first Category I code of its kind in this series' timeline. Compare it with the 2023 baseline, where billable AI lived mostly in Category III codes and in a handful of procedures. A permanent code with a defined payment rate converts a pilot budget line into a product line.

📊 Data point: CPT 75577 took effect January 1, 2026 at a hospital outpatient rate of $950.50, per AuntMinnie.

For other specialties, the lesson is about sequencing. Clearance gets a product to market. Evidence gets it a code. A code gets it paid. The 2023 claims study showed that only a few procedures had completed that sequence. The 2026 code set shows at least one has.

What this means for health system leaders, FQHCs, and AI diagnostics

For health system leaders, the three-year record says to separate three questions that vendors blend. Is the tool authorized? Is it reimbursed? Has it been tested on our patients? Most clinical AI in the 2023 to 2026 data passes the first test, a few pass the second, and the third is the one almost nobody on the independent side of the divide can answer.

For FQHCs, the data is a warning and an opening. The ASTP/ONC numbers show independent facilities using predictive AI at 37% and checking accuracy locally at 41%. Community health centers need vendors who bring validation data, monitoring, and a reimbursement path, not just a license. A reimbursed model changes the economics: it is how a tool reaches a clinic that cannot fund a pilot.

For radiology and AI diagnostics, the device list is the easy part. Radiology still holds roughly three-quarters of authorized devices, and the Category I code for coronary plaque shows how a diagnostic can reach permanent payment. In lung cancer screening, where earlier detection is the point, the same path applies: clear it, prove it, get it coded, and make it work in the facilities that have the fewest resources. Oatmeal Health's LungAI is investigational and not FDA cleared.

The record from 2023 to 2026 supports a simple conclusion. Adoption is no longer the hard part. Proof, access, and local validation are.

Key References

Wu K, Wu E, Theodorou B, Liang W, Mack C, Glass L, Sun J, Zou J. "Characterizing the Clinical Adoption of Medical AI Devices through U.S. Insurance Claims." NEJM AI 2024;1(1), published November 9, 2023. DOI: 10.1056/AIoa2300030.

Zink A, Boone C, Joynt Maddox KE, Neprash HT, et al. "Artificial Intelligence in Medicare: Utilization, Spending, and Access to AI-Enabled Clinical Software." The American Journal of Managed Care, Health IT special issue, May 31, 2024. Everson J, Nong P, Richwine C. "Uptake of Generative AI Integrated With Electronic Health Records in US Hospitals." JAMA Network Open, December 12, 2025.

U.S. FDA, "Artificial Intelligence-Enabled Medical Devices," page current as of September 22, 2026. MedTech Dive (950 devices as of August 2024). AuntMinnie, FDA July 2025 update. OncoDaily, August 18, 2026. PharmaDossier (annual and cumulative counts from the Innolitics 510(k) tracker).

American Medical Association, "AI usage among doctors doubles as confidence in technology grows," March 12, 2026; Fierce Healthcare coverage of the same survey.

ASTP/ONC Data Brief 80, "Hospital Trends in the Use, Evaluation, and Governance of Predictive AI, 2023 to 2024" (September 2025); AHA Center for Health Innovation, November 4, 2025; Healthcare Dive, September 18, 2025; Healthcare Brew, November 24, 2025. KGOU/StateImpact Oklahoma, August 19, 2026 (Brian Whitacre, Oklahoma State University).

Rock Health, 2025 year-end and H1 2026 funding overviews; Healthcare Dive and Fierce Healthcare coverage of H1 2026 (July 2026). Menlo Ventures, "2025: The State of AI in Healthcare," October 21, 2025. AuntMinnie, "CMS finalizes Medicare reimbursement for Caristo's coronary plaque analysis." MedCentral, "CPT 2026 Updates Expand Lab, Category III, and AI Codes"; AMA, "CPT Category III Codes Long Descriptors" (updated October 2, 2026); AMA, "CPT codes offer the language to report AI-enabled health services" (October 10, 2025); TechTarget, "AMA releases CPT 2026 code set, adds codes for health AI," September 11, 2025; Medikode, "AMA September 2026 CPT Panel: AI Coding Rules Take Shape."

Jonathan Govette is the Co-Founder and CEO of Oatmeal Health, which is building AI lung cancer diagnostics to catch cancers earlier in underserved communities. He writes The Oatmeal Bite for health system leaders, FQHCs, and radiology teams.

Read that again.

Not AIDS-defining illnesses. Not opportunistic infections. Lung cancer. A disease with a proven, covered, widely recommended early-detection tool. A disease where screening reduces mortality by approximately 20%. A disease that, if caught early, is survivable.

And yet a new study published in the International Journal of STD & AIDS — covered by The American Journal of Managed Care in October 2026 — found that fewer than 3 in 10 eligible patients living with HIV ever received a low-dose CT scan during a six-year study window.

That is not a guideline failure. The guidelines improved substantially in 2021.

That is not a coverage failure. Low-dose CT screening is covered for qualifying patients.

That is not a contact failure. These patients were already engaged in care, regularly seeing providers for antiretroviral therapy management.

That is an execution failure. And executio…

Read more

Tri-State is not a small footnote. It was founded in 1987 as a nonprofit federally qualified health center and served 16,085 patients in 2025. Of those, 70.8 percent lived at or below 200 percent of the federal poverty level. These are people for whom a locked clinic door does not mean a rescheduled appointment. It means a gap in insulin, blood pressure medication, or a cancer follow-up that may not be filled anywhere else within driving distance.

For health system executives and FQHC leaders watching from outside, the instinct may be to treat this as an isolated event - a single organization with a particular problem. That i…

Read more

In September 2026, AstraZeneca presented the 8-year exploratory overall survival (OS) update from the Phase III ADAURA trial at the World Conference on Lung Cancer. The data were published simultaneously in the Journal of Thoracic Oncology. The numbers, already striking at three and five years, are now definitive: adjuvant osimertinib (Tagrisso) reduces the risk of death by 47 percent in patients with resected Stage II-IIIA EGFR-mutated non-small cell lung cancer. This is the longest survival follow-up ever reported from a global Phase III trial in this setting.

For thoracic oncologists, radiation oncologists, and health system leaders, the question is no longer whether osimertinib belongs in the adjuvant setting. That question is closed. The question is whether your multidisciplinary workflow is built to actually deliver the benefit these numbers …

Read more

On Thursday, September 24, 2026, patients arrived at Tri-State Community Health Centers in Berkeley Springs, West Virginia, and found the doors locked. A handwritten sign said the organization was facing an “organizational emergency.” Four more sites across Maryland and Pennsylvania had gone dark the same way, for a total of five, with little or no advance notice.

Tri-State is not a small footnote. It was founded in 1987 as a nonprofit federally qualified health center and served 16,085 patients in 2025. Of those, 70.8 percent lived at or below 200 percent of the federal poverty level. These are people for whom a locked clinic door does not mean a rescheduled appointment. It means a gap in insulin, blood pressure medication, or a cancer follow-up that may not be filled anywhere else within driving distance.

This piece looks at what is publicly known about the closure, how HRSA oversight is supposed to work, and why the financial and administrative structure of the health center program …

Read more

760,000 people just lost their ACA coverage in one week.

The government calls it a fraud crackdown. And by the numbers, it is. CMS canceled approximately 315,000 unauthorized enrollments in the Federal Health Insurance Marketplace, covering more than 760,000 individuals. The agency expects to recover roughly $2.2 billion in taxpayer-funded subsidies.

Those are extraordinary numbers. The largest single enforcement action against Marketplace fraud in recent memory.

But the number that keeps me thinking isn't the $2.2 billion. It's the 760,000 people who woke up last week without health insurance and had to figure out what just happened to them.

That's the story worth unpacking.

Read more

Because that is exactly what is happening right now.

This month, a coalition of 67 health care industry and advocacy groups sent a letter to congressional leadership calling the No Surprises Act arbitration process a "massive unforeseen crisis in employer-sponsored coverage." Those words belong to Melissa Bartlett, senior vice president for health policy at the ERISA Industry Committee. She is not a fringe voice. She speaks for the people managing the benefits your employees depend on.

Here is how a law designed to protect patients became a $15 billion revenue engine for a handful of provider groups.

What the No Surprises Act Actually Did

The No Surprises Act, passed in 2020, did something genuinely good. It shielded patients from getting blindsided by out-of-network bills after emergency visits or procedures at in-network facilities. Before the law, a patient could walk into an in-network hospital, get treated by a…

Read more

The FDA just funded a jury of AI judges for radiology reports.

Not one AI. A panel of them. Cross-examining each other.

On September 16, 2026, the FDA awarded Cognita Imaging a $1.29 million research contract to build what the company calls an "LLMs-as-a-jury" evaluation framework for autonomous, generative AI radiology reports. The 18-month contract was funded under the FDA's Broad Agency Announcement program for advanced research and development in regulatory science, effective June 22, 2026.

This is not a product clearance. It is the FDA funding the infrastructure to evaluate a whole new category of AI before that category overwhelms the agency's ability to review it. That distinction matters enormously to anyone building, buying, or deploying radiology AI.

Read more

Texas hospitals were bleeding $27 million a day. That just ended.

On September 17, 2026, Governor Greg Abbott announced the restoration of nearly $12 billion in Medicaid supplemental funding to Texas hospitals, ending what federal officials confirmed was the first halt to this funding program since it launched in 2014. The pause began September 1, the start of Texas's new fiscal year, and lasted just over two weeks. But the damage to hospital balance sheets - and to trust in the stability of American healthcare finance - is not erased by a wire transfer.

This is not just a Texas story. The underlying dispute - over how states structure provider taxes to access federal Medicaid matching dollars - is now the center of a national rulemaking fight that could permanently reshape how every state funds hospital care for low-income patients. What happened in Texas is a preview. The main event is still coming.

1. What Happened and Why the Freeze Started

The mechanics of Texas's Medicaid supplement…

Read more