Behavioral-health payments are changing. Care coordination is a different problem.

A new economic architecture for behavioral-health payment is taking shape across Medicare and Medicaid this year: an inpatient psychiatric rate update, a proposed physician pay cut, an expanded community-clinic financing model, and rules meant to align Medicare and Medicaid insurance for dual-eligible beneficiaries. None of it, on its own, says whether care gets coordinated for the patients who cross both programs to get it.

A study identifies where the two programs overlap in paying for behavioral-health services, raising a separate question about how care is coordinated. Dr. Hyunjee Kim, a health economist at Oregon Health and Science University, found that 30.6 percent of Dual-Eligible Special Needs Plan enrollees who used a behavioral-health service in 2021 had that care paid for by both Medicare and Medicaid, not just one. That figure measures where the money came from, not whether the patient’s care was coordinated. Kim’s own reading of what it might mean is more specific: patients and families who cross both programs’ rules and provider networks may face added navigation and coordination burdens that a payment-rate change does not, by itself, resolve.

Her study is specifically about D-SNP enrollees, a Medicare Advantage subset of the broader dual-eligible population, using 2021 claims data. Among the roughly 493,000 of those enrollees who had a claim with a primary behavioral-health diagnosis that year, 62.0 percent used only Medicare-paid services and 7.4 percent used only Medicaid-paid services; the remaining 30.6 percent, shown below, used services paid by both programs.

That 30.6 percent figure is not itself proof that anyone in it experienced fragmented care. Kim said the group “may be at higher risk of experiencing care fragmentation because they or their family members have to navigate two separate systems with different rules, provider networks, etc.” and because, in her words, Medicare and Medicaid “often don’t talk to each other in a timely manner.” She offered a concrete illustration of how that could play out, not a measured result of her study: a dual-eligible enrollee receives acute psychiatric treatment paid by Medicare, is discharged home, and then needs ongoing community-based behavioral health services paid by Medicaid. The hospital and the community provider, she said, “likely operate separately and do not share information in real time,” which can produce delayed follow-up and gaps in care during exactly that transition.

Asked directly whether recent CMS behavioral-health payment changes, in either the physician fee schedule or the inpatient psychiatric facility rule, address that specific coordination problem, Kim was unambiguous: “As far as I know, those changes are not aimed at improving care coordination between Medicare and Medicaid for dual-eligible individuals.” That is the throughline for everything below: payment-rate provisions and insurance-integration requirements are advancing on different tracks. The IPF rate update, the outlier cap, and the physician fee schedule’s conversion-factor cut and coding provisions are payment-mechanism changes, not integration measures, and none of them was built to close the coordination gap Kim describes. That does not mean no federal effort exists on the integration side; it does, in the dual-eligible insurance rules discussed further below. The two tracks are simply not the same project.

What Medicare finalized for psychiatric hospitals

CMS’s final rule updating Medicare’s Inpatient Psychiatric Facility Prospective Payment System for fiscal year 2027, issued July 29, sets the per diem base rate at $912.40, up from $892.87. That is roughly a 2.3 percent payment-rate update, a 3.2 percent market-basket increase minus a 0.9 percentage point productivity adjustment, which CMS also calls its estimated $60 million aggregate impact system-wide. The base rate itself moved a smaller 2.2 percent in dollar terms, because a separate wage-index budget-neutrality factor of 0.9989 is applied on top of the market-basket update. Neither figure describes what any single facility actually collects, which also depends on case mix and wage-index geography.

A second, unrelated provision does not move with the annual update at all: starting FY 2028, effective October 1, 2027, facility-level outlier payments are capped at 20 percent of an IPF’s total Medicare PPS payments annually, for facilities with 50 or more stays a year. CMS delayed the cap a year and lowered its originally proposed stay threshold after commenters warned it could disrupt access to care for high-acuity patients. For operators with material Medicare psychiatric-inpatient exposure, such as Acadia Healthcare and Universal Health Services, whether the FY 2028 cap applies depends on facility-level outlier-payment history this piece has no visibility into. The cap matters to qualifying facilities whose outlier payments would otherwise exceed 20 percent of their total Medicare PPS payments; that is a dated, facility-specific constraint taking effect in FY 2028, not a quantifiable company-wide effect that follows from the rule alone.

What Medicare has only proposed for outpatient behavioral health

The outpatient-side rule is not yet final. CMS’s Calendar Year 2027 Medicare Physician Fee Schedule remains proposed as of this writing; CMS’s own regulatory docket shows no final rule issued, the comment period closed September 14, and CMS has historically finalized this rule in late October or November, as it did for CY 2026 on October 31, 2025. As proposed, both statutory conversion factors fall for 2027, roughly 1.19 percent for clinicians in a qualifying Advanced Alternative Payment Model and 1.68 percent for everyone else, because the one-time 2.5 percent bump Congress provided only for 2026 does not carry forward. That is a system-wide cut with no behavioral-health carve-out, and among the behavioral-health-relevant provisions this piece covers, it is the one figure still genuinely unresolved: CMS can adjust it before finalizing. The rule contains other provisions outside behavioral health that also remain open before finalization; the conversion factor is simply the one that matters most here, not the rule’s only unsettled piece.

Two other behavioral-health-relevant provisions inside the same rule are already settled, through mechanisms this rulemaking does not control, and they work differently from each other. The waiver removing the in-person visit requirement for Medicare mental-health telehealth visits, through December 31, 2027, is statutory: Congress set that deadline in the Consolidated Appropriations Act, 2026, and the CY 2027 rule only makes conforming regulatory changes to match a law already in force. It preserves a delivery option, letting a mental-health telehealth visit continue without a preceding in-person visit, and has no bearing on the conversion-factor cut, which is calculated separately. The exemption of time-based codes, the codes psychiatric evaluation, psychotherapy, and most behavioral-health integration billing run on, from a separate 2.5 percent efficiency reduction applied to procedural services, is a different kind of provision: CMS finalized it in the CY 2026 rule and it continues automatically. Unlike the telehealth waiver, this exemption is itself a payment adjustment, determining which codes avoid an additional reduction layered on the conversion-factor cut, which narrows relative impact on this specialty compared to others without a comparable exemption. Neither provision changes the size of the conversion-factor cut itself.

A different financing model, not just more of it

Medicaid’s Certified Community Behavioral Health Clinic program, operating since 2017, is a structurally different kind of payment change already underway, separate from Medicare’s annual-update mechanics. What changed in 2026 is its reach, not the model itself: HHS and SAMHSA announced on May 28 that 10 more states were joining the federal Section 223 CCBHC Demonstration, bringing it to 30 participating states, and that 31 of the 46 states with any CCBHC now support the model through Medicaid, via the Demonstration or a state plan amendment. Two further SAMHSA mechanisms, an accreditation overhaul for expansion-grant clinics and a $12 million state-planning-grant program, are advancing that count further this year; neither has produced a final decision yet.

The structural difference is what a CCBHC prospective payment system replaces, and what it does not guarantee. Ordinary Medicaid behavioral-health reimbursement pays fee-for-service, triggered by each billed encounter. A CCBHC PPS, run under one of four CMS-defined methodologies, two daily rates and two monthly, still requires at least one qualifying encounter to trigger payment, but it changes the unit of payment from procedure to package: the rate is cost-based, calibrated to a clinic’s expected or projected cost of delivering the full required service package, built from its own submitted cost reports and periodically rebased. It is explicitly not a true-up to actual cost: CMS’s own guidance states that reconciling the PPS rate to a clinic’s actual cost “is not allowed” under the Demonstration, so a clinic that spends less than projected keeps the difference, and one that spends more gets no automatic top-up before the next rebasing. The two monthly methodologies, PPS-2 and PPS-4, require outlier payments and quality bonus payments tied to SAMHSA’s quality measures; the two daily methodologies, PPS-1 and PPS-3, permit optional quality bonus payments instead, with no outlier-payment provision in CMS’s own methodology descriptions, and PPS-3 additionally requires its own separate rates for qualifying crisis services. That is an accountability model built around a defined scope of care and a projected cost estimate, not a guarantee of full cost recovery or of any individual clinic’s financial viability, which still depends on how its actual costs move against that estimate before the next rebasing.

Same insurer is not the same thing as one system

Some integration effort does exist, aimed specifically at the dual-eligible population, and Kim’s own account of it carries a distinction worth preserving exactly. Federal rules already let dual-eligible enrollees receive both Medicare and Medicaid behavioral health benefits through a single insurer, structured as a Fully Integrated Dual Eligible Special Needs Plan or a Highly Integrated Dual Eligible Special Needs Plan. The two are not interchangeable: a FIDE SNP’s capitated Medicaid contract must cover behavioral health by default, alongside primary, acute, and long-term-care benefits, while a HIDE SNP’s contract only has to cover long-term care or behavioral health, meaning a plan can qualify as highly integrated while still carving behavioral health out entirely to Medicaid fee-for-service. As of the most recent published breakdown, drawn from 2023 enrollment data and now several years stale, roughly 8 percent of D-SNP enrollees are in a FIDE SNP and 35 percent in a HIDE SNP, out of a national dual-eligible population of about 12 million.

Kim’s own proposed fix does not stop at “put both programs under one insurer.” She thinks an even better model is one where a dual-eligible enrollee receives all Medicare and Medicaid benefits, behavioral health included, through a single plan, instead of through the Medicare and Medicaid sides of a single insurance company that can still operate separately.

“Even when a single insurer offers both Medicare and Medicaid coverage, there is still a risk that the Medicare and Medicaid side of the single insurance company operate separately and do not coordinate care effectively. In contrast, when a single plan is accountable for all Medicare and Medicaid benefits, there is a greater chance for care coordination and therefore reduced gaps in communications across Medicare/Medicaid providers and care settings.”

— Hyunjee Kim, corresponding author, JAMA Network Open dual-eligible behavioral health study

That distinction, common corporate ownership against unified plan-level accountability, is not a hypothetical concern on Kim’s part. It is closer to the current evidence base than a plain reading of “FIDE and HIDE SNPs exist” would suggest. A systematic review published in JAMA Health Forum in July 2024 examined care-coordination outcomes across integrated dual-eligible models and found no measurable difference in care-coordination processes in eight of the nine studies it covered; for the subgroup of beneficiaries with serious mental illness specifically, it found the evidence insufficient to characterize a direction at all. A more recent JAMA Network Open study, published in June 2026, documents that integrated D-SNP enrollment among dual-eligible beneficiaries with serious mental illness rose from 2.4 percent in 2019 to 17.2 percent in 2022, and states plainly that, to its authors’ knowledge, no study has yet examined how that enrollment relates to actual care or outcomes for that population. Integrated-plan enrollment has expanded, but the studies cited here do not establish whether these arrangements improve behavioral-health coordination or outcomes for beneficiaries with serious mental illness.

One further CMS change, already finalized, not merely proposed, pushes in the direction Kim’s stronger model implies without yet reaching it. Under 42 CFR 422.514(h), where a Medicare Advantage organization, its parent, or a co-owned affiliate holds a Medicaid managed-care contract enrolling full-benefit dual eligibles in the same service area as its D-SNP, that D-SNP is limited to one plan for full-benefit duals in that area, and the alignment phases in instead of applying immediately: from 2027 through 2029, only new enrollment is restricted to people already enrolled in, or actively enrolling in, the affiliated Medicaid plan, with a harder requirement that existing unaligned members transition starting in 2030. CMS carved out exceptions, including state-approved geographic carve-outs, continued coordination-only plans for partial-benefit duals, and a crosswalk exception for transitioning members. That tightens the link between the Medicare and Medicaid sides of a single company’s offering, on a multi-year timetable with defined exceptions, not an immediate blanket requirement. It does not, on its own, create the single-plan-accountable model Kim described; it makes the same-insurer version of integration more common, not the harder version she thinks solves the coordination problem.

Where reimbursement outruns the people who deliver it

A payment rate is a necessary condition for care to happen, not a sufficient one. HRSA’s most current quarterly data, current as of June 30, 2026, counts 7,109 designated Mental Health Health Professional Shortage Areas nationally, covering more than 157 million people, with only 26.53 percent of the estimated need met, a worse position than the December 2025 figures most secondary reporting still cites. HRSA’s own workforce projections put the 2038 psychiatrist shortfall at 43,810 full-time-equivalent providers on a current-use basis, alongside a projected shortfall of nearly 100,000 psychologists that rises past 136,000 once unmet need, not only current utilization, is counted.

Federal workforce incentives moved in a mixed direction this cycle. The National Health Service Corps’ main loan-repayment program increased its award for primary-care disciplines to $75,000 for a full-time commitment, but explicitly excluded behavioral and mental health providers from that increase, leaving their award at the prior $50,000 level. The one behavioral-health-specific track that did see a meaningful increase this cycle, the NHSC Substance Use Disorder Workforce Loan Repayment Program, rose to a base of $75,000, reaching $80,000 only with an optional language-proficiency add-on, a narrower, addiction-specific slice of the behavioral health workforce, not a general increase. A CCBHC PPS can pay a clinic its full projected cost for a comprehensive service package, and a psychiatric hospital can collect a higher per diem rate, in a market where this year’s targeted federal workforce dollars did not go to the general behavioral-health workforce those payment structures assume will exist.

What this adds up to for operators, payers, and investors

Read separately, each of these tracks supports only a narrow claim. Read together: payment-rate provisions, the IPF update, the outlier cap, and the physician fee schedule’s conversion-factor cut, were never built to address Medicare-Medicaid coordination, and none of them do. The D-SNP alignment requirement was built for exactly that purpose, and its behavioral-health-specific effect simply has not been demonstrated yet.

For psychiatric facility operators, the annual rate update and the FY 2028 outlier cap are separate exposures to model: a modest system-wide rate increase now, and a facility-specific ceiling later that matters only to facilities whose outlier payments would otherwise exceed 20 percent of their total Medicare PPS payments, which this piece cannot quantify for any single company. For outpatient behavioral-health providers, digital platforms included, the conversion-factor cut is a live payment number that will not be final for weeks; the telehealth waiver that lets them keep billing without an in-person visit first is a separate, already-settled delivery question the finalization does not touch. For CCBHC operators and the states funding them, prospective payment changes the unit of reimbursement from encounter to package and rewards comprehensive care delivery over volume, but it is not a guarantee that the projected rate covers any given clinic’s actual cost or ensures its financial viability. For managed-care organizations selling into the dual-eligible market, aligning the Medicare and Medicaid sides of an insurance company under 42 CFR 422.514(h) is a distinct achievement from demonstrating that alignment improves behavioral-health coordination, which the studies examined here have not yet confirmed. And for any service-intensive delivery model, psychiatric hospitals and psychotherapy practices among them, a higher rate or a more favorable payment structure assumes a workforce to deliver the care, and federal data shows that capacity tightening, not loosening, while this year’s targeted incentive dollars flowed narrowly.

What to watch

The CY 2027 Physician Fee Schedule final rule, expected around late October or November, will settle the actual size of the conversion-factor cut; the telehealth waiver and the time-based code exemption are not what that finalization decides. SAMHSA’s still-open decisions on CCBHC accreditation, and the outcome of this year’s state planning-grant competition, will determine which states build PPS infrastructure next. The FY 2028 IPF outlier cap becomes a live financial event for high-cost facilities on October 1, 2027. The D-SNP alignment requirement’s new-enrollment restriction begins the same cycle, with the harder alignment of existing members still years out, in 2030. And the more consequential, quieter marker is whether a study closing that gap gets published: direct evidence on whether integrated dual-eligible plans improve behavioral-health outcomes for beneficiaries with serious mental illness, which the JAMA Network Open study’s own authors say does not yet exist to their knowledge, addressing the same coordination problem the alignment requirement and its predecessor SNP structures were built for.