Back to Insights
Health Policy

CMS's New ABA Toolkit Meets Three Different Medicaid Models: Louisiana, Mississippi, and Arkansas

August 14, 2026

CMS's new toolkit documents Medicaid autism-therapy spending outpacing the population it serves, with $198.4 million in improper payments already found nationwide. Louisiana, Mississippi, and Arkansas each administer ABA through a different structure — so the toolkit lands differently in each state.

CMS's new toolkit documents Medicaid autism-therapy spending outpacing the population it serves, with $198.4 million in improper payments already found nationwide. Louisiana, Mississippi, and Arkansas each administer Applied Behavior Analysis through a different structure, so the toolkit lands differently in each state.

Medicaid and CHIP spending on Applied Behavior Analysis for children with autism grew 421 percent between 2021 and 2025, while the number of children actually receiving those services grew 189 percent. That gap is why the Centers for Medicare & Medicaid Services treats ABA oversight as urgent rather than routine in the toolkit it released on August 4: a 170-page playbook built from a review of more than 260 literature sources and over 240 state Medicaid policy documents. Its premise: ABA, the leading evidence-based therapy for autism spectrum disorder (ASD), has grown faster in cost than in reach, and seven state audits have already found $198.4 million in improper Medicaid payments tied to autism services. HHS Secretary Robert F. Kennedy, Jr. framed the release as a tool for states "to identify bad actors, protect families, and hold providers accountable"; CMS Administrator Dr. Mehmet Oz put it more bluntly: every dollar lost to fraud or waste is a dollar that never reaches a child who needs it. Neither framing disputes that ABA, delivered by a trusted provider under a properly supervised, individualized treatment plan, remains one of the more effective interventions available for young children with ASD.

The toolkit is written for a wide audience: state Medicaid and CHIP agencies, the managed care organizations that administer benefits on their behalf, the full range of licensed behavior analysts and technicians who deliver ABA, and the caregivers and patients who depend on the system working as designed. Louisiana, Mississippi, and Arkansas give three distinct tests of how that audience responds, because none of the three runs ABA the same way.

The number CMS opened with

The toolkit's own data tell a more precise story than the topline percentages suggest. Total Medicaid and CHIP ABA spending rose from roughly $1.94 billion in 2021 to $10.1 billion in 2025. Over that same span, the number of children with an ASD diagnosis receiving any Medicaid or CHIP service grew 67 percent, from 1.15 million to 1.92 million. Within that group, the smaller subset who received ABA specifically grew 189 percent, while what Medicaid paid for their ABA grew 421 percent. Spending is compounding roughly twice as fast as the population it serves, and that arithmetic is the toolkit's central justification for a heavier emphasis on utilization management, documentation, and provider oversight.

Seven states have already had that arithmetic tested by federal auditors. HHS-OIG reviews of ABA and related autism-services billing have so far found:

State HHS-OIG Report Confirmed Improper Payments Recommended Federal Refund
Colorado A-09-24-02004 (Feb. 2026) $77.8 million N/A
Indiana A-09-22-02002 (Dec. 2024) $56.5 million ($76.7M potential) $39.4 million
Maine A-01-24-00006 (Jan. 2026) $45.6 million ($22.4M potential) $28.7 million
Wisconsin A-06-23-01002 (July 2025) $18.5 million ($94.3M potential) $12.2 million
Massachusetts, Nebraska, Nevada State/OIG reviews Included in totals below N/A

Across all seven states, auditors reported a combined $198.4 million in improper payments, driven by thin session-note documentation, billing for non-therapy or non-billable time, services delivered by unqualified or under-supervised staff, and weak post-payment review. None of the three states in this region has appeared on that list yet. That is a different thing from saying any of them has been checked and cleared.

Seersucker take: For any Medicaid program running ABA at scale, the audit pattern above is a preview, not a warning specific to one state. Medicaid agencies, MCOs and PASSEs, and the ABA and behavioral health provider organizations serving their members should treat "we haven't been audited yet" as a planning window, not a clean bill of health, and use it to pressure-test session notes, supervision logs, and medical-necessity documentation against the exact failure modes HHS-OIG has now catalogued seven times over.

Three states, three different delivery systems

Louisiana, Mississippi, and Arkansas each built a different administrative structure for ABA long before this toolkit existed, and the toolkit's recommendations land on each one differently as a result.

Louisiana Medicaid delivers ABA entirely through the five Healthy Louisiana managed care organizations: Aetna Better Health, AmeriHealth Caritas, Healthy Blue, Humana Healthy Horizons, and Louisiana Healthcare Connections. Coverage begins with a Comprehensive Diagnostic Evaluation from a qualifying professional, moves through MCO-network provider selection, and requires plan approval before services start; authorizations run no longer than 180 days before a reassessment is due.

Arkansas runs a hybrid. Most Medicaid beneficiaries ages 18 months to 21 with an ASD diagnosis get ABA reviewed on a fee-for-service basis, with Acentra Health acting as the state's utilization-review contractor: it requires physician documentation of medical necessity, a BCBA review of the request, and escalation to a BCBA-D or physician reviewer when a denial is on the table. Beneficiaries enrolled in a PASSE (Arkansas's provider-led shared-savings entities for higher-acuity behavioral health and developmental-disability populations) or in the state's Autism Waiver go through their PASSE care coordinator instead. Arkansas is also the one state the toolkit itself singles out as actively revising its rules: a footnote notes the state anticipates updating its ABA Medicaid manual, based on extensive discussion with its ABA stakeholder group, to shift observation-threshold requirements from a per-provider caseload basis to a per-beneficiary basis.

Mississippi delivers ABA through its MississippiCAN managed care plans under the EPSDT benefit for enrollees under 21. The ABA provider, not the family, submits the treatment plan and carries the burden of demonstrating medical necessity to the plan. Mississippi also just built new state-level infrastructure for autism specifically: House Bill 807, passed in the 2025 legislative session, created a Division of Autism Services within the Department of Mental Health, giving the state a dedicated body that did not exist when most of the toolkit's source policy documents were collected.

State Delivery system Prior authorization Supervision requirement In CMS's rate comparison
Louisiana Fully capitated managed care, 5 MCOs CDE, then MCO approval; 180-day authorization, reassessed every 6 months CaBA/RBT to LBA at 2:10 (2014 Medicaid Services Manual) No, absent from both rate tables
Arkansas Hybrid: FFS reviewed by Acentra, PASSE for higher-acuity members Physician medical necessity, BCBA review, escalation for denials Minimum 5% treatment-hour observation plus 1 hour every 30 days; rule under revision Yes, appears in the state rate table
Mississippi Managed care, MississippiCAN plans Provider-submitted treatment plan, MCO medical-necessity review State licensure law requires supervision by a licensed behavior analyst or psychologist; no toolkit-reported ratio Yes, holds the nationally highest reported rate for CPT 97152

Seersucker take: A capitated model like Louisiana's puts program integrity largely in the MCOs' hands, subject to state contract oversight. A hybrid model like Arkansas's puts it more directly on a state utilization-review contractor, with a separate track for the PASSE population. That difference should shape where each state's Medicaid agency and provider associations focus their compliance investment over the next year, rather than assuming a single national playbook applies evenly.

Where the region sits on the toolkit's own supervision bar

The toolkit's clinical guidance calls for one to two hours of BCBA-level supervision for every ten hours of direct treatment (1-2:10). Louisiana's requirement, unchanged since 2014, sits at 2:10, the top of that range rather than the middle. Arkansas's minimum observation threshold, 5 percent of treatment hours plus one hour every 30 days, is harder to translate directly into the same ratio, but it is the one supervision framework in the entire toolkit that the state itself is already in the process of revising, based on the stakeholder feedback CMS cites. Mississippi's statute requires licensed supervision but does not specify a ratio the toolkit reports, which leaves the state without a clean national benchmark on this particular measure.

That leaves Louisiana with the least room to respond to the toolkit by simply tightening a number, since its number is already near the ceiling of what CMS recommends. It leaves Arkansas already mid-conversation with its own provider community about how supervision should be measured, a conversation state agencies and provider associations elsewhere may want to watch. And it leaves Mississippi's new Division of Autism Services building rules at a moment when a specific, defensible supervision ratio is exactly the kind of policy the toolkit will make providers and legislators ask for.

The rate table two of three states are on

CMS's toolkit includes a state-by-state comparison of 2025 fee-for-service rates for the core ABA billing codes, drawn from the Council of Autism Service Providers' 50-state benefit survey and cross-checked against actual Medicaid claims data. Mississippi and Arkansas both appear in it. Mississippi, in fact, holds the nationally highest reported rate for CPT 97152 (behavior identification supporting assessment by a technician) at $41.74, more than four times West Virginia's low of $9.90. Louisiana's row is blank throughout.

The toolkit doesn't explain that omission. A blank cell means CASP either didn't receive a reportable Louisiana rate or CMS couldn't verify one against claims data, not that no rate exists. But the absence matters on its own terms: as more states and advocacy organizations use this table to benchmark provider reimbursement, a state that isn't in it is a state that providers, legislators, and advocates cannot easily compare to its neighbors. For a program run entirely through five MCOs whose rate-setting isn't uniformly public to begin with, that is a transparency gap layered on top of an already fragmented rate picture, and it is a gap Louisiana shares with none of the states next door.

Seersucker take: Provider associations negotiating MCO or PASSE contracts in any of the three states have an opening here. Mississippi's providers can point to a federally published, top-of-market rate on at least one code. Arkansas's providers have a documented baseline. Louisiana's providers are negotiating against a state that has no external, federally verified number to be held to, which cuts both ways depending on which side of the table you're sitting.

What it means for Medicaid agencies, MCOs, providers, and families

State Medicaid agencies in all three states should expect this toolkit to become the reference document the next legislative session, provider association, or HHS-OIG auditor reaches for first. Louisiana's LDH will likely face questions about why the state's rates aren't independently verifiable; Arkansas's DHS has a live opportunity to finish its supervision-rule revision in a way that anticipates the toolkit's recommendations rather than reacting to them after the fact; Mississippi's new Division of Autism Services is essentially building its rulebook at the same moment CMS published one.

Managed care organizations and PASSEs carry the front-line compliance burden in every state covered here. The toolkit's call to "align managed care oversight with FFS program integrity safeguards" applies directly to Louisiana's five MCOs and Mississippi's MississippiCAN plans, and applies with a twist to Arkansas's PASSEs, which already sit outside the standard Acentra review pathway and will need their own documentation standard to match.

ABA providers and the broader behavioral health provider community should read the toolkit as a documentation warning, not a clinical one. Every HHS-OIG finding to date traces back to thin session notes, ambiguous billing of supervision time, or care delivered without a matching, current individualized treatment plan, not to disputes over whether ABA itself works. Agencies in all three states should expect tighter prior-authorization documentation requirements and prepayment claims edits in the coming year.

Caregivers and patients in Louisiana and Mississippi, where a CDE- or treatment-plan-first authorization model already resembles the toolkit's template, are unlikely to see the front door of the process change much. Mississippi families should note one distinction that has nothing to do with the toolkit and everything to do with which coverage their child has: the state's private insurance autism mandate, which applies to fully insured commercial plans, caps ABA coverage at age 8 (extendable with documented medical necessity) and 25 hours per week. Medicaid, governed by federal EPSDT, carries no such fixed ceiling for medically necessary care to children under 21. A Mississippi child covered by a parent's employer plan and a Mississippi child covered by Medicaid can face materially different limits for the same diagnosis, and that gap is easy to miss if a family assumes their commercial plan works the way Medicaid does.

The bigger picture

A federal government that has just spent a year cataloguing where ABA oversight failed in seven other states is not likely to wait long before asking whether the eighth, ninth, and tenth are next, and a region running three different delivery models gives CMS three different places to look. Arkansas already has a live stakeholder process revising its supervision rule. Mississippi already has a new Division of Autism Services standing up its own rulebook. Louisiana has neither in motion, and it is also the one state of the three with no independently verified rate on record. None of that means Louisiana, Mississippi, or Arkansas has done anything wrong; it means each state now has a specific, different piece of unfinished business to close before a federal auditor, rather than a legislative session, decides the timeline.

Seersucker Strategies works with healthcare providers, Medicaid agencies, and associations across Louisiana, Mississippi, and Arkansas navigating state Medicaid policy, MCO and PASSE contracts, and program integrity requirements. Reach out if this toolkit affects your organization's compliance posture or reimbursement strategy.

Sources: CMS, State Medicaid & CHIP Applied Behavior Analysis Toolkit (August 2026); CMS Newsroom, "CMS Launches New State Toolkit to Protect Children with Autism, Strengthen Oversight of Applied Behavior Analysis Services" (Aug. 4, 2026); HHS-OIG Reports A-09-24-02004 (Colorado), A-09-22-02002 (Indiana), A-01-24-00006 (Maine), A-06-23-01002 (Wisconsin); Louisiana Department of Health, Arranging Medicaid-Covered Applied Behavior Analysis fact sheet; Louisiana Department of Health, Applied Behavior Analysis, Chapter Four of the Medicaid Services Manual (2014); Arkansas Department of Human Services, Applied Behavior Analysis Services Provider Manual; Acentra Health, Arkansas ABA utilization review program description; Mississippi Division of Medicaid, EPSDT/ABA coverage guidance; Mississippi Insurance Department, autism mandate summary; Mississippi Department of Mental Health, Division of Autism Services (created by 2025 H.B. 807); Council of Autism Service Providers, 50-State Applied Behavior Analysis Medicaid Benefit Comparison (2025).