Two Warrants, Six Federal Statutes, No Charges Yet
ABA Centers of America is the subject of a federal criminal investigation, according to warrant applications reported by Behavioral Health Business on September 3. The investigators are the HHS Office of Inspector General and the Department of Justice’s criminal fraud section. The applications were first surfaced by the newsletter Reinforcement & Ruin.
The timeline in the documents: in February 2025, investigators obtained a warrant for the Microsoft account data of founder, chairman, and former CEO Christopher Barnett. Using that material plus interviews with employees, health plan officials, and parents, they went back for more. A second application, filed under seal at the end of 2025, sought the digital records of officials at ABA Centers of America, its corporate parent ICBD Holdings, and several sister companies. BHB puts the new targets at 25 additional executives and employees.
The warrants cite six federal offenses: false statements relating to health care matters, wire fraud, health care fraud, conspiracy to commit health care or wire fraud, conspiracy to commit money laundering, and transacting in criminal proceeds.
The scale investigators describe is hard to read past. Across the payer data they collected, ABA Centers and its entities allegedly billed health plans $2.4 billion and were reimbursed more than $480 million. Cigna commercial plans alone were billed $119 million from 2022 to 2024 and paid $26.5 million.
Now the part that matters most. No one has been charged. As BHB notes, nothing in the applications has been presented as specific charges against any individual, and there has been no determination of guilt. A warrant application is an investigator’s argument for probable cause, not a finding. The company declined to comment to BHB. Everything below is an allegation.
Why This Lands Differently After June
In June we covered Optum cutting off ABA Centers of America: Optum-managed families were discharged June 5, and neither side said why. That post laid out the civil fights with Point32Health and Publix.
BHB’s reporting doesn’t connect the investigation to the Optum decision, and we won’t claim it does. As of this writing, we haven’t found any public statement from Optum explaining its decision. But one detail stands out to us: the warrant documents say executives labeled a set of insurers “super payors,” and the list begins with OptumHealth Behavioral Solutions and United Behavioral Health, followed by BlueCross BlueShield of Massachusetts, Cigna, Anthem, and Harvard Pilgrim/Tufts. Payer special investigation units talk to federal investigators. The Cigna SIU is quoted in these very documents.
The civil cases are also moving toward the criminal frame. BHB reports that a Florida federal judge recently found the Publix allegations could also have been brought as a criminal action.
The Allegations Are an Auditor’s Checklist
Strip out the jet (a $9.95 million aircraft allegedly bought through a Barnett-owned leasing company) and the Red Sox suites, and what’s left is a list of billing behaviors. Each one is something payer analytics and auditors already test for, at practices of every size.
That’s the useful way to read this story. You don’t need to be accused of anything for these tests to run against your claims. The question is whether your records answer them before anyone asks.
Here’s each alleged scheme, what a clean practice’s records look like, and how you’d prove it.
1. Location: The Claim Says Where the Session Happened
The allegation: investigators say the company would change the location where services took place to get higher reimbursement. Reinforcement & Ruin’s reading of the affidavits describes services delivered in Massachusetts billed through New Hampshire offices, where rates were higher.
What clean looks like: the service location and place of service on the claim match where the session physically occurred. Home sessions bill as home. Clinic sessions bill as the clinic that hosted them. A practice with offices in two states or two rate regions never routes a session to the better-paying office after the fact.
How to prove it: place of service captured on the session record at the time of service, not assigned at claim creation. Session notes that name the setting. Staff location data (mileage logs, schedules, clock-ins) that agrees with the billed location. Run one report this month: sessions where the billed location differs from the scheduled location. It should be zero, or every exception should have a note. If you just adopted telehealth, check the POS and modifier rules too; we covered those in our telehealth billing post.
2. Credentialing: Network Status Changes for Real Reasons
The allegation: de-credentialing providers so their services could be billed out-of-network at higher rates.
What clean looks like: a clinician’s payer enrollment status is stable and boring. When someone leaves a network, there’s a business reason on file (a contract ended, a payer closed a panel), and it applies to the practice, not to one clinician’s caseload being moved to a higher rate.
How to prove it: a credentialing history for each rendering provider showing enrollment dates, terminations, and the reason for each change. Clients whose rendering provider changed should have a clinical or scheduling reason documented, not a payer reason. If your rate per unit for the same client jumped with no change in the service, you’d want to know why before an SIU does.
3. Only Clinical Services Get Billed
The allegation: billing for services that were not clinical, including sending clinicians to bowling alleys with families. Cigna’s SIU told investigators the company’s medical records lacked basic information to justify bills, such as the names of treating clinicians or the time increments provided.
What clean looks like: community outings can be legitimate ABA when they target goals in the treatment plan: generalization, tolerance, safety skills. The difference is the paper. A billed session in the community has targets, data, and a note that says what was taught and measured.
How to prove it: every billed unit ties to a session note with the rendering clinician, start and stop times, the goals targeted, and data collected. Cigna’s complaint is the minimum bar, and it’s a low one. If a note would read the same with the clinical content deleted, it’s not billable. This is the same discipline as defensible RBT supervision: the service you did but can’t prove is the service you pay back.
4. Coding: A Human Can Explain Every Code
The allegation: automated upcoding built into the company’s homegrown EMR.
What clean looks like: software speeds up coding but never decides it. The code on the claim follows from what the note documents: who delivered the service, what kind of service it was, how long it ran. Rules that map sessions to codes are visible, reviewed, and don’t quietly favor the higher-paying option.
How to prove it: a code distribution report by payer and by clinician, compared against your client mix. If one code’s share jumps after a system change or a new payer contract, a person should be able to explain the jump clinically. Claim edits run before submission, and when someone overrides one, there’s a record of who and why. An audit trail on the claim itself is the difference between “our system did it” and “here’s exactly what happened.”
5. Hours Follow the Assessment, Not the Insurance Card
The allegation: a separate clinical assessment team for patients with “super payor” coverage, and denying treatment unless parents agreed to high-hour plans or single-case agreements. Reinforcement & Ruin reports the assessment team prescribed 35 to 40 hours weekly regardless of clinical need.
What clean looks like: one assessment process for every client. Recommended hours come from the assessment, the treatment plan, and the clinical rationale, with ranges that vary by child. Your average recommended hours shouldn’t move with the payer.
How to prove it: pull recommended weekly hours by payer. If your commercial clients average 38 hours and your Medicaid clients average 22 with similar clinical profiles, that’s a pattern an auditor will describe for you. The fix isn’t lowering hours for kids who need them. It’s being able to show each recommendation came from the child’s assessment and outcome data, which is also how you defend hours when payers push back.
6. Collect the Cost Share You Report
The allegation: waiving patient cost-sharing, and then threatening to collect previously unpaid cost shares if patients tried to reduce hours or stop treatment. The documents argue that waived cost-sharing removes a patient’s incentive to check that the provider isn’t billing for unnecessary services or services not rendered. The civil counterclaim we covered in June described $1-a-month payment plans.
What clean looks like: a written financial policy. Copays, coinsurance, and deductibles get billed and collected. Hardship reductions exist, but they’re case by case, based on documented financial need, and never advertised as a reason to choose you. Collection practice never depends on whether a family wants fewer hours.
How to prove it: patient statements that reflect the patient responsibility on each remittance, collection history per family, and a file for every hardship adjustment showing the need and who approved it. Run a report of families with large patient balances and little or no collection activity. Each one is either a hardship file or a problem.
Run the Self-Audit Before Anyone Else Does
You can run all six tests in a single afternoon with reports most practice systems already produce:
- Location mismatches: billed location vs. scheduled and documented location, last 90 days.
- Credentialing changes: every rendering provider whose network status changed in the last 12 months, with the reason.
- Note completeness: a random sample of 20 billed sessions, checked for clinician name, times, goals, and data.
- Code distribution: share of each CPT code by payer and clinician, compared against six months ago.
- Hours by payer: average recommended weekly hours grouped by payer.
- Cost-share collection: patient responsibility billed vs. collected, by family.
Anything you find, fix it and write down what you changed. A documented correction reads very differently to an auditor than a pattern nobody noticed.
This is also cheap insurance against the scenario we wrote about in July: Medicaid payment holds that close practices before any fraud is proven. Allegations get built from exactly these patterns. And after the Minnesota autism fraud indictments, no one should assume a federal investigation is somebody else’s problem. And the CMS ABA toolkit released in August already tells state Medicaid agencies how to tighten claims review and flag fraud, so expect more of these tests, not fewer.
Where Software Fits, Honestly
No system makes a practice honest. What a system can do is make the six tests above easy to run and the answers easy to produce.
In StreamABA, place of service is recorded on the session and carries through to the billing line, so the claim reflects where care happened. Every claim traces back to a session note with a full audit trail. Claims run through validation and NCCI edits before submission. Credential expiration alerts and authorization tracking flag problems before they reach a claim, and patient statements and payment plans keep cost-share collection on the record instead of in someone’s head.
The records you’d want in front of an investigator are the same records you need to run a clean practice on an ordinary Tuesday.
Want help running this six-point self-audit against your own claims? Reach out at [email protected].
Follow us on LinkedIn and X for more industry updates.
Sources
- Behavioral Health Business: ABA Centers of America at the Center of Federal Criminal Investigation (September 3, 2026)
- Reinforcement & Ruin: The Warrants, Part 6 of the investigation in ABACOA (September 2, 2026)
- Behavioral Health Business: ABA Centers of America Lays Off Staff (June 12, 2026)
- Breaking News ABA: Optum Cuts Off ABA Centers of America (June 22, 2026)
