CMS Hands States a 174-Page ABA Playbook — and It Is Not About Clinical Care

The toolkit creates no new federal rules. It tells every state Medicaid agency what good oversight is supposed to look like.

On 4 August 2026, CMS issued a 174-page Applied Behavior Analysis Toolkit for state Medicaid and CHIP agencies. Read the headline and you might relax: the document does not establish any new federal requirements and stops short of imposing a standardized national framework for ABA.

That is precisely why it is worth your attention. A toolkit is not a rule — it is a statement of what the federal government considers good practice, handed to the regulators who write the rules that actually bind you.

What CMS is pointing states toward

  • Provider enrollment and credentialing. States are encouraged to revisit who is permitted to bill and what they must hold to do so.
  • Utilization management. Explicit direction to build the machinery for spotting improper billing.
  • Documentation red flags. The toolkit names them: overlapping treatment, excessive service hours, and reevaluations that do not reflect progress.
  • Telehealth. States are asked to limit virtual delivery as a primary vehicle for ABA, on fraud-risk grounds.
  • Measurable outcomes. Repeated emphasis on demonstrating that children are actually improving.

The number behind it

ABA spending in Medicaid and CHIP rose 421% between 2021 and 2025. Growth at that rate attracts scrutiny regardless of whether the care behind it was appropriate, and this document is what scrutiny looks like when it becomes policy guidance.

What you must know or do

No rule in your state changed today. What changed is that your state now holds a federal description of good oversight — and the red flags it names are answerable against your own files this week.

Take the three the toolkit names and read them as questions:

  • Overlapping treatment. Do any two authorized services land in the same clock hour for the same child?
  • Excessive service hours. For your highest-intensity clients, is there a written clinical rationale for that intensity, dated at the time it was set rather than reconstructed later?
  • Reevaluations that do not reflect progress. Pull your last ten. Do they show movement against a baseline, or restate the goals from the previous authorization period?

If you deliver virtually, take the telehealth line first. Every other item here is a paperwork problem. That one is a service-line problem — it goes to how care is delivered, not how it is recorded.

North Carolina rewrote its policy three days before this toolkit appeared. That is your preview of how far a state goes once it moves.