CMS Requires More Restrictive Definition of Medical Frailty in New Medicaid Work Requirements Rule
On June 1, 2026, the Centers for Medicare and Medicaid Services (CMS) issued a long-anticipated interim final rule that will guide state implementation of Medicaid work requirements. The 2025 reconciliation law requires 43 states to condition Medicaid eligibility for adults in the Affordable Care Act (ACA) Medicaid expansion group and enrollees in partial expansion waiver programs (Georgia and Wisconsin) on meeting work requirements starting January 1, 2027 or sooner at state option. Most Medicaid adults are already working or face barriers to work, and the Congressional Budget Office estimates that the new requirements will not meaningfully increase employment by Medicaid enrollees, but that millions of people will end up uninsured.
At nearly 400 pages, the rule is quite long and complex and will require time to fully assess its implications. Given the complexity of the provisions in the rule, states will likely face significant challenges in operationalizing the requirements in the next six months. With the short implementation timeline, states were already making implementation decisions that may not align with the rule. These challenges present risks for states given the potential for audits and financial penalties, but also for individuals who now face greater barriers to obtaining and maintaining coverage.
Significantly, the rule adopts a restrictive definition of medical frailty that differs from states’ early expectations. States have been eagerly awaiting clarification on how to define medical frailty. Early indications from CMS offered through informal meetings with states hinted that the federal definition might mirror an existing medical frailty definition used for determining individuals who are exempt from receiving an alternative benefit package, and that states would be given flexibility to go beyond the federal definition. However, on both counts, the rule adopted a more restrictive approach, first tying medical frailty specifically to the ability to comply with the community engagement requirement (i.e., the ability to work) and prohibiting states from adding categories of individuals to the medical frailty definition. To operationalize this definition, the rule requires states to develop lists of health conditions but prohibits them from categorically exempting people with those conditions without considering an individual’s ability to meet the community engagement requirement based on their health. The rule does not provide states with a list of diagnoses or criteria for measuring severity or ability to meet new requirements. For example, the rule does not allow states to exempt all people with cancer, HIV, Parkinson’s disease, or multiple sclerosis. This significant change may pose challenges for states that had been moving forward with developing lists of diagnosis codes assuming a less restrictive definition and could make it more difficult for individuals to document medical frailty.
The rule temporarily allows for self-declaration in limited circumstances to verify work or exemption status, including for medical frailty. Acknowledging that states will need time to establish systems upgrades to access new data sources and particularly to implement the medical frailty exemption, the rule allows states to accept self-declaration of work or exemption status when reliable data are not available through 2027. However, starting January 2028, except for verifying medical frailty, states must request documentation from individuals when the state does not have data on file. In the case of medical frailty, states are permitted to accept self-declaration only one time during the enrollment period. For example, if self-declaration of medical frailty is accepted at application, individuals must provide documentation at renewal, typically six months later.
While a temporary period of allowing people to self-attest to their compliance with work requirements will ease the effects for some, the more restrictive definition of medical frailty adopted by the Trump Administration will be challenging for states to implement and could lead to more people falling through the cracks and losing coverage.
Jennifer Tolbert
Director of State Health Policy and Data
Deputy Director Program on Medicaid & Uninsured
A comment on the other half than the one you addressed in the post, of the total package of damage to the ACA done last year. The non-extension of the ACA expanded subsidies, that already went into effect of Jan 1, 2026.
As you have caught in a 5/19/2026 KFF report, https://www.kff.org/affordable-care-act/what-we-know-so-far-about-2026-aca-marketplace-enrollment-premiums-and-deductibles/
it's looking like 3 to 5 million, of about 24 million covered last year on the ACA exchanges, will wind up losing coverage.
(There is also a very good NY Times story, from a few weeks before the KFF report, showing similar numbers, from similar sources:
https://www.nytimes.com/2026/05/01/business/obamacare-enrollment-decline.html )
What I wanted to focus on here, is that it looks like CMS has already not released data on ACA coverage numbers that it has released in the past, apparently to avoid showing the large coverage drops.
(So, if it were unemployment data or inflation data that is regularly released, a thousand economists would point out the absence of data.
In this case, policy-wonky Charles Gaba caught it, and from reading his Substack, I am now aware.
I don’t know if anyone else is aware! Probably someone at KFF somewhere is aware. I hope someone over there is aware!)
As well (caught a bit in your KFF report), leaked information exists that the large coverage losses have become evident to CMS internally. And that CMS is planning a misleading propagandistic response, claiming that the losses are due mainly to the administration's successful crackdown on ACA fraud.
--
So, on the ACA expanded subsidies that were allowed to lapse 1/1/26, data has been starting to roll in from non-federal sources showing something like 3 to 5 million people, of 24 million who had coverage on the ACA exchanges in 2025, losing coverage. (As in the KFF report, and the Times story, both linked to prior.)
This information available to the public is part of the new data that continuously rolls in from non-federal non-CMS sources, as initial very limited open-enrollment-period only data, from CMS, which counts people auto-renewed as having coverage, even if they don't wind up paying the premium, gets superseded. (That limited initial data showed only a 4.9% enrollment drop for those 24 million.)
In any case, after releasing that initial open-enrollment-period data for 2026 showing only a 4.9% drop in enrollment, with the big defect in the data that people who wind not paying the premium are not counted as dropping coverage, I don't believe anything has been released by CMS (Dr. Oz's Center for Medicaid and Medicare Services, which also manages the ACA).
The much larger coverage-drop numbers indicated in the KFF and NY Times reports, counting people not paying premiums as dropping, have come from other sources. (State exchanges, KFF surveys, and a major consulting company report with access to insurance company data.)
So, has the CMS been blocking the release of information on the coverage drops?
Charles Gaba actually caught one case where the data, usually presented as part of a report on Medicare and Medicaid and CHIP, was omitted.
Charles reports on that here: https://charlesgaba.substack.com/p/cms-posts-january-2026-medicaid-chip
Quoting Charles:
"Until now, the summary report also included a brief mention of total effectuated ACA marketplace enrollment in Qualified Health Plans (QHPs), rounded off to the nearest 100,000. As of December effectuated enrollment was ~21.8 million people.
However, starting this month, this data point is missing...and that’s not by accident; it includes this footnote:
As of the January 2026 data, Marketplace enrollment data are no longer included in this report but will be available separately soon."
So, we may already have a case of data usually made available being suppressed.
Otherwise, on the data within CMS, I don't know if they usually release it and are holding it back, but we know they have it. Further, they appear to be preparing to falsely claim the coverage drops are due to reduction of fraud on the exchange.
The source is this recent NOTUS story by former Washington Post Reporter Paige Winfield Cunningham, which post seems to access leaked information from people inside of CMS:
https://www.notus.org/healthcare/aca-healthcare-dropped-insurance-numbers-subsidies
Quoting that:
"More than one in five people who enrolled in health insurance through http://healthcare.gov/ during open enrollment and in the weeks immediately following were dropped from coverage for failing to pay their first month’s premium, according to internal Centers for Medicare and Medicaid Services, or CMS, documents obtained by NOTUS that haven’t been made public.
The roughly 21% decline in enrollment in the 30 states using the federal marketplace is significantly higher than the rate of last year, when 12% of enrollees dropped off over the same time frame.
The numbers support widespread fears that the end of extra, pandemic-era subsidies, which congressional Republicans declined to extend in December, would leave Affordable Care Act plans unaffordable for some Americans.
Faced with such a stark drop in enrollment, leadership at CMS, which is led by Administrator Mehmet Oz, is seeking to attribute a majority of the enrollment declines to rooting out fraud rather than people not paying their premiums, according to three CMS sources. The sources said it’s unlikely fraud is behind most of the cancellations."
(I and other readers of Charles Gaba's substack were pointed to the leaked information in NOTUS by his post: https://charlesgaba.substack.com/p/breaking-cms-admits-over-30-million )
--
Misleading Propagandistic Response: I think I see it coming from a pair of participants:
I think there is starting to be evidence of the ground being laid for the erroneous claims coming that the coverage drops are due to the administration stopping fraud on the ACA exchanges, emanating from Dr. Oz and the Paragon Health Institute. To save space, for those interested, it,s in a comment elsewhere, this one:
https://econjared.substack.com/p/why-are-people-so-damn-mad/comment/261555717
Thank you for reporting that here on your KFF Substack.
(I caught the information your are reporting on in a New York Times article of yesterday
https://www.nytimes.com/2026/06/01/upshot/trump-medicaid-work-requirements.html
and have already posted a link to that article in a few Substack comment sections that I know people interested in the ACA and expanded Medicaid frequent.)
In context, for those not up on the details, the issue you are pointing out is part of the One Big Beautiful Bill (OBBB) cuts, timed to be after the midterm election and start beginning of 2027.
They are a part of the total package of Republican cuts from both OBBB and the Republican's not allowing the extension of the ACA expanded subsidies, that the Democrats shut down the government at the end of last year in an unsuccessful attempt to try and undo.
(The latter non-extension-of-expanded-subsidy cuts have occurred already, and went into effect 1/1/26, and affect the people on the on-exchange half of the ACA. I have a separate comment coming here on the effects of that.)
The OBBB cuts which you have discussed an aspect of in your post, affect mostly people on the expanded-Medicaid half of the ACA, rather than the on-exchange half. As I indicated, they are timed for start of 2027, to be after the midterm election.
I, myself, predict massive technical foul ups when the OBBB expanded Medicaid cuts start, since 51 separate state and D.C. Medicaid agencies, many under-resourced and not that technically adept, have to deal with new work requirements on expanded Medicaid, and more frequent (twice a year minimum) expanded Medicaid eligibility checks.
I predict, because of being technically overwhelmed, numerous states will wind up not giving coverage to large numbers of people who, under the new rules, are supposed to have coverage.
Each state will create its own custom-version of a nightmare for numerous people, as it messes up things in its own custom way!
(This observation about technical foul-ups is not abstract. I relied on the ACA and expanded Medicaid system here in Massachusetts for a period of 8 years, and I had about 5 major foul-ups that put me in primal fear for my finances and/or my health outcome each time! I've been on Medicare-for-Me with a medigap for a few years, and so, in the intervening time, perhaps the state agencies have gotten a bit better. OBBB has now thrown in extensive complications in my state, and all the rest. Creating a likely nightmare.)
---
Between both the non-extension of the ACA expanded subsidies, and the OBBB requirements, the country has taken a step backwards on its health insurance system.
Unfortunately, I see no way politically possible way to start undoing the damage until Jan 20, 2029 at the earliest. In that case, only if all federal elections go the Democrats way.