Key Takeaways
- Managed care now controls most Medicaid behavioral health spending. As of July 2025, 42 states contract with Medicaid MCOs, and most enrollees receive their behavioral health benefits through them, either directly or through a carved-out behavioral health organization.
- Network adequacy is largely on paper. A 2025 OIG data brief found that on average less than a third of the behavioral health workforce was included in MCO provider networks, and in many plans more than a third of listed providers were inactive.
- Federal Medicaid funding is contracting. H.R.1, signed in July 2025, cuts roughly $1 trillion from Medicaid over a decade, with downstream pressure on the rates and contracts that determine MCO behavioral health spending.
- The carve-out debate is back. After a decade of states moving toward integrated, “carve-in” arrangements, advocates in New York and elsewhere are arguing the experiment has failed for behavioral health and pushing to carve services back out.
Most of the policy debate about Medicaid takes place at a level of abstraction that obscures how the program actually works for the more than 70 million people it covers. Behavioral health makes that abstraction harder to sustain. The question of whether someone with a serious mental illness can find a psychiatrist who takes their plan, or whether a parent can get their child into therapy within 30 days, does not turn on the size of the federal match or the political fortunes of expansion. It turns on a managed care contract: which provider networks the plan has built, what rates the plan pays, what utilization controls the plan applies, and what the plan’s parent firm calculates it can afford given its capitation rate. Forty-two states now contract with Medicaid managed care organizations, and as of 2024, the five largest publicly traded firms (Centene, UnitedHealth Group, Elevance, Molina, and CVS Health/Aetna) accounted for roughly 47 percent of all Medicaid MCO enrollment. What those companies decide about behavioral health, often through actuarial inputs and contract negotiations the public never sees, increasingly determines what access looks like.
That has always been true to some extent. What has changed in the past two years is the visibility of the gap between MCO behavioral health networks and the populations those networks are supposed to serve, and the speed at which the federal financing environment is contracting around plans that already struggle to meet existing standards.
How Medicaid managed care now delivers most behavioral health care
The structure varies, but the pattern is consistent. As of July 2025, all states except Alaska, Connecticut, Maine, Vermont, and Wyoming operate some form of Medicaid managed care. Forty-two states contract with comprehensive MCOs; 13 operate primary care case management programs; 28 contract with one or more limited-benefit prepaid health plans for services like behavioral health, dental, or transportation. Idaho, following passage of HB 345, expects to wind down its PCCM program by December 2025 and stand up comprehensive MCOs by January 2029. Nevada plans to expand its MCOs statewide in 2026.
For behavioral health, two structural questions matter most. The first is whether services are “carved in” or “carved out” of the comprehensive MCO contract. In a carve-in arrangement, the MCO is responsible for both physical and behavioral health benefits, and a single capitated rate covers both. In a carve-out, the state contracts separately with a managed behavioral health organization (often a Beacon, Carelon, or Optum subsidiary) that administers behavioral health benefits and bears separate financial risk. Carve-outs were common in the 1990s and 2000s; the past decade has seen a steady migration toward carve-in arrangements, in part because of 2016 federal regulations that incentivized integrated financing and in part because of the administrative burden of running separate parity-compliance regimes.
By 2019, only six states (California, Colorado, Maryland, Michigan, Pennsylvania, and Utah) maintained statewide carve-outs of all behavioral health benefits. The rest of the carve-out activity now happens at the plan level: comprehensive MCOs subcontract behavioral health benefit management to a specialty BHO, creating what researchers call a “secondary carve-out.” For providers, the practical effect is the same. They credential and contract separately with the BHO, follow its authorization workflows, and bill it for behavioral health services, even though the underlying state contract is technically integrated.
The second structural question is what the MCO is paid. Medicaid managed care rates are set by state actuaries on a per-member, per-month basis, must be actuarially sound, and are reviewed and approved by CMS. To limit administrative spend, federal rules require states to develop capitation rates targeting a medical loss ratio of at least 85 percent. As of July 2025, 33 MCO states reported they always require remittance payments from plans that fail to meet the MLR threshold, which is the principal mechanism for clawing back funds plans collected but did not spend on care. The remaining states either require remittance only sometimes or not at all.
What plans do with that capitated spend is, within MLR and contract constraints, up to them. The MCO decides what providers to contract with, what to pay them, what authorizations to require, and what utilization patterns to flag. Those decisions, more than any state legislative debate, are what determines who gets behavioral health care under Medicaid in 2026. Acuity has documented a parallel dynamic in the in-network migration reshaping the SUD industry, where addiction treatment providers face the same MCO-driven contracting logic.
What the OIG found in Medicaid managed care behavioral health networks
In October 2025, the Department of Health and Human Services Office of Inspector General published a data brief that has not received the editorial attention it deserves. OIG analysts reviewed 60 plans across 10 counties, including five urban and five rural counties in five states. They focused on three categories of behavioral health providers: psychiatrists, psychologists, and clinical social workers. The findings, which echo themes Acuity has covered on the dual-diagnosis billing and regulatory silos still built into behavioral health, should be alarming to anyone who reads them in full.
Three-quarters of the Medicare Advantage plans reviewed had less than 25 percent of the county’s behavioral health workforce in their networks. Half of Medicaid managed care plans fell into the same “limited network” category. That means enrollees in those plans may not have access to 75 percent or more of the behavioral health providers in their county. Among the providers listed in network directories, on average 55 percent of those in Medicare Advantage plans and 28 percent of those in Medicaid managed care plans were “inactive,” meaning they had not provided a single service to a plan enrollee in the review period. Seventy-two percent of those inactive providers, by OIG’s analysis, should not have been listed at all: the practice address was wrong, the provider had retired, the provider had said they would not see plan enrollees.
The phenomenon, known as a “ghost network,” is not new to anyone in the field. What is new is the federal validation. OIG’s recommendations to CMS were structurally modest, namely to use data more aggressively to monitor networks, to work with states to improve directory accuracy in Medicaid managed care, and to continue exploring a national directory. CMS did not explicitly concur or non-concur but indicated steps it had already taken or planned. State-level enforcement has been more aggressive. On February 19, 2026, New York Attorney General Letitia James announced a settlement with EmblemHealth requiring the insurer to pay $2.5 million in penalties and fees, plus restitution to members, after the AG’s office found EmblemHealth’s directories overstated behavioral health provider availability by as much as 80 percent. Similar investigations are active or threatened in several other states.
The reasons providers cite for staying out of MCO networks are familiar to any clinician who has tried to bill them. In the OIG’s provider survey, the two most-cited deterrents to network participation were administrative burden and low payment rates. Both are within the MCO’s control. Both are precisely what the MCO is incentivized to manage tightly under a capitated contract.
What Medicaid MCOs themselves say they are doing about behavioral health access
A 2025 study published in The Milbank Quarterly offered a rare inside view of MCO strategies on behavioral health access. Researchers led by Jane Zhu at Oregon Health & Science University interviewed 27 administrators and executives across 19 Medicaid MCO carriers operating at local, regional, and national scales. The findings are substantive but should be read against the OIG’s network-adequacy backdrop.
MCOs interviewed said they perceived acute access challenges in three areas: care for children and adolescents; rural geographies; and crisis and transitional services. To address these, they reported a common bundle of strategies: contracting with a “core group” of Medicaid-focused behavioral health providers to deliver a substantial share of care; outreach, training, and workforce support programs; rate enhancements above the contracted floor; telehealth and mobile-unit care models; and “high-touch case management” to align members with appropriate providers and levels of service.
Several findings cut against the conventional framing of MCO behavioral health practice. Plans operating in carve-out states reported a wide variation in how much they knew about their members’ behavioral health needs and utilization. Some had longstanding collaborations with community-based behavioral health plans tied to social services, including housing, food, and parenting supports. Others had “virtually no knowledge” of their members’ behavioral health utilization at all, the structural consequence of a carve-out arrangement that intentionally separated those data streams. The implication is that whether an MCO is a useful access lever depends substantially on the financing architecture the state has chosen, not just on the plan’s clinical operations.
State context shaped the strategies in other ways. States that contracted Medicaid services regionally, with one plan per region, allowed MCOs to focus on building deep workforce relationships in a defined geography. States with multiple plans operating in overlapping regions sometimes saw plans competing for the same providers, which Zhu and colleagues note can complicate workforce-development partnerships. Rate-setting practices varied widely, as did administrative and regulatory requirements imposed at the state level. The lesson, the researchers concluded, is that MCO behavioral health performance is heavily contingent on the contract terms and policy context the state establishes.
The behavioral health carve-out debate, reopened by New York
After a decade in which the policy consensus moved toward carve-in financing, the carve-out question has resurfaced in 2025 and 2026. The most visible challenge has come in New York, where the state moved behavioral health benefits into mainstream Medicaid managed care in 2015 under Governor Andrew Cuomo. In December 2025, Lauri Cole, executive director of the New York State Council for Community Behavioral Healthcare, published a Times Union commentary calling the experiment a “disaster” and urging Governor Kathy Hochul to carve behavioral health back out. The argument, echoed by The Alliance for Rights and Recovery and other advocacy groups, is that insurers have been paid handsomely while failing to comply with parity laws and contract provisions, and that funding meant for community-based providers has been siphoned off through denials, delays, and inaccurate provider networks. A 2022 New York AG investigation cited by the advocates found that 86 percent of providers listed in mental health network directories at 13 health plans were misleading or “ghost” providers.
The empirical literature on the question is mixed. A 2021 study of carve-in versus carve-out arrangements in the Portland, Oregon tri-county area found that carving in behavioral health services to comprehensive MCOs was associated with improved access, particularly for Black beneficiaries. Other research has found that carve-out arrangements can preserve specialty provider networks and protect funding for clinically complex populations, including individuals with serious mental illness. A widely cited 2022 review in Health Affairs Forefront concluded that carve-in and carve-out models can have comparable performance if designed to facilitate clinical and financial integration, but that “comparable performance” is not what either arrangement is consistently delivering in practice.
Several states have adopted hybrid approaches that try to capture the advantages of both. Arizona and Florida have focused integration efforts primarily on individuals with serious mental illness, while leaving general-population behavioral health under different arrangements. New York carved behavioral health into mainstream MCOs but designated a subset of plans, called HARPs, specifically for individuals with SMI. The variation underscores the absence of a single right answer; what it does not change is that the mechanics of behavioral health access for Medicaid enrollees are, in 2026, increasingly determined by managed-care contracting rather than by direct state administration.
What H.R.1 does to Medicaid managed care behavioral health
Into this environment came the One Big Beautiful Bill Act, which President Trump signed into law on July 4, 2025. The legislation cuts federal Medicaid funding by approximately $1 trillion over ten years, the largest reduction in the program’s 60-year history. The Congressional Budget Office estimates that roughly 10 to 12 million people will lose Medicaid coverage as a result of the law’s eligibility, work-requirement, and redetermination provisions, phased in across late 2026 and into 2027.
Several provisions affect behavioral health access through MCO contracting. Beginning at the end of December 2026, all states must conduct Medicaid eligibility redeterminations for the expansion population every six months rather than annually, an administrative change that is expected to push some eligible enrollees off the rolls. Beginning January 1, 2027, most expansion adults will be subject to 80-hour-per-month “community engagement” or work requirements, which the bill’s critics, including the AMA and APA, say will result in further eligible-but-disenrolled populations. The enhanced FMAP for Medicaid expansion sunsets December 31, 2026, removing a federal incentive that had supported coverage of low-income adults in many states. The bill caps state directed payments at 100 percent of Medicare rates for expansion states and 110 percent for non-expansion states beginning in 2028, which a KFF analysis warns will tighten the rate-floor that supports specialty provider participation, including behavioral health.
For behavioral health specifically, the law contains some narrow exemptions. The new cost-sharing requirements that take effect in October 2028 (up to $35 per service for expansion enrollees with incomes between 100 and 138 percent of the federal poverty level) exempt primary care, mental health, and substance use disorder services, as well as services delivered by FQHCs, behavioral health clinics, and rural health clinics. The exemption is meaningful, but it operates against a backdrop of overall coverage contraction that will shrink the population eligible for those services in the first place. The Rural Health Transformation Program, which appropriates $50 billion over five years for rural providers, will offset only a fraction of the rural Medicaid spending reductions estimated by KFF.
Plans are already responding. In recent quarterly earnings calls, the largest publicly traded MCO parent firms reported that the rate updates they had received from states were not keeping pace with the acuity and utilization patterns of their post-redetermination membership, and that they expected this misalignment to extend through 2025. Behavioral health utilization was specifically cited as a driver of higher costs, alongside pharmaceuticals and long-term services and supports. As capitation rates and acuity diverge, plans face pressure to manage utilization more tightly, including through prior authorization, network design, and case management decisions that affect behavioral health access. Acuity has covered the related compliance reckoning splitting the ABA and behavioral health M&A market in two, a downstream effect of the same financing pressures.
What the policy environment is leaving in place for state Medicaid behavioral health rates
The implications for providers (particularly the safety-net providers, FQHCs, and CCBHCs that serve much of the Medicaid behavioral health population) are not abstract. KFF’s 50-state Medicaid budget survey for fiscal years 2025 and 2026 found that the number of states planning behavioral health rate increases dropped from 23 in FY 2025 to 14 in FY 2026, a notable contraction. State Medicaid programs are facing a more tenuous fiscal climate; the same KFF survey found that six states implemented provider rate restrictions in FY 2025, and another six are planning them for FY 2026, compared with one and two states respectively in FY 2024 and FY 2023.
The combination is what should worry behavioral health observers. Federal funding is contracting. State rate setting is tightening. Federal parity enforcement is on hold. MCO networks are documented to be limited and inaccurate. The carve-out debate is reopening because the carve-in experiment has not delivered what its proponents promised in many states. And the underlying behavioral health workforce, which Acuity has documented in a related feature on burnout and reimbursement, is leaving insurance panels at rates that the existing pipeline cannot replace.
None of this is hidden. The OIG report is public; the H.R.1 implementation timeline is public; the KFF surveys are public; the parity enforcement statement is public. What is harder to see is the cumulative effect, because it lives inside thousands of individual contract negotiations, network-build decisions, and prior-authorization workflows that no single regulator monitors in full. For Medicaid enrollees with behavioral health needs, the consequence is that access is increasingly determined not by what state law says they are entitled to, but by what their MCO’s contract, network, and utilization-management practices actually deliver. For policymakers, that is a fact that the next round of state Medicaid budgets and CMS rule-makings will not be able to avoid. Acuity’s ongoing regulatory coverage tracks the state and federal actions reshaping the answer.
Frequently Asked Questions
How many states use Medicaid managed care for behavioral health services in 2026?
As of July 2025, all states except Alaska, Connecticut, Maine, Vermont, and Wyoming operate some form of Medicaid managed care. Forty-two states contract with comprehensive Medicaid MCOs, and 28 states contract with one or more limited-benefit prepaid health plans, which often include behavioral health. Most Medicaid behavioral health spending now flows through MCOs or their behavioral health subcontractors.
What is the difference between a behavioral health carve-in and a carve-out arrangement?
In a carve-in arrangement, the comprehensive Medicaid MCO covers both physical and behavioral health services under a single capitated contract. In a carve-out, the state contracts separately with a managed behavioral health organization that administers behavioral health benefits. As of 2019, only six states (California, Colorado, Maryland, Michigan, Pennsylvania, and Utah) maintained statewide carve-outs of all behavioral health benefits. Many MCOs subcontract behavioral health management to specialty BHOs, creating de facto plan-level carve-outs.
What is a Medicaid behavioral health “ghost network”?
A ghost network is a provider directory that lists clinicians as in-network when they are not actually available, because they have retired, changed practice, declined to see plan enrollees, or are not accepting new patients. An October 2025 OIG data brief found that on average 28 percent of providers listed in Medicaid managed care behavioral health networks were inactive, and 72 percent of those inactive providers, in OIG’s analysis, should not have been listed at all.
How will H.R.1 affect Medicaid behavioral health coverage?
The One Big Beautiful Bill Act, signed July 4, 2025, cuts federal Medicaid funding by approximately $1 trillion over ten years. CBO estimates that 10 to 12 million people will lose Medicaid coverage between late 2026 and 2027 due to work requirements, more frequent eligibility redeterminations, and the sunset of enhanced FMAP for expansion. Cost-sharing requirements that take effect in October 2028 exempt mental health, substance use, and primary care services, but the overall coverage contraction will reduce the population accessing those services.
How are MCOs trying to expand Medicaid behavioral health access?
A 2025 study in The Milbank Quarterly interviewed 27 executives at 19 Medicaid MCOs and identified five common strategies: contracting with a core group of Medicaid-focused behavioral health providers; outreach, training, and workforce support programs; rate enhancements above the contracted floor; telehealth and mobile-unit care models; and high-touch case management to align members with appropriate care levels. Strategies were highly dependent on the state’s rate-setting practices and contract requirements.
Why are some states reconsidering behavioral health carve-outs?
Advocates in New York and elsewhere argue that integrating behavioral health into mainstream Medicaid MCOs has resulted in inadequate networks, low reimbursement, parity violations, and funding diverted from community-based providers to plan administration and profit. The empirical literature is mixed: some studies find carve-in arrangements improve access for certain populations, while others find carve-outs better preserve specialty networks for individuals with serious mental illness. Several states use hybrid models, including specialized plans for SMI populations, to try to capture the advantages of both. Acuity’s coverage of state Medicaid ABA reimbursement reform tracks parallel structural debates within behavioral health’s autism services segment.
