Marcus Autism Center Study Finds 91% Drop in Severe Behavior

What the Study Found

ATLANTA – Children who entered Marcus Autism Center’s intensive outpatient program because their behavior put themselves or others at immediate risk saw a 91% reduction in those behaviors by the end of treatment, according to a retrospective review published online June 30, 2026, in the Journal of Developmental & Behavioral Pediatrics (JDBP).

The research team from the Marcus Autism Center and the Emory University School of Medicine reviewed electronic health records for 40 patients aged 6 to 19 who were treated between January 2020 and December 2023. All showed severe externalizing behaviors, including aggression, disruption, and self-injury. Aggression appeared in 87.5% of patients, disruption in 82.5%, and self-injury in 67.5%. Every patient had an autism diagnosis, and 80% also had an intellectual disability. The study calls this group intellectual and developmental disability, or IDD.

When they finished the program, 82.5% of patients had also reached their main treatment goal, which was set together with their families.

“Findings provide provisional evidence regarding the benefits of intensive behavioral treatment for children with IDD and severe externalizing behavior.” – McMahon et al., Journal of Developmental & Behavioral Pediatrics (2026)

The authors are clear about the study’s limits: it examined only one program, relied on a retrospective record review, and included only 40 patients. Lead author Meara McMahon and her co-authors at Marcus and Emory, including Mindy Scheithauer, Nathan Call, William Sharp, and Colin Muething, do not make stronger claims and describe the findings as provisional evidence.

The Model Behind the Number

Patients attended the program Monday through Friday, five hours a day, for about five months. The intervention was structured and tailored to each person using applied behavior analysis principles. This schedule totals about 25 clinic hours per week, or around 500 hours over a typical admission. This is much more intensive than standard outpatient ABA.

Marcus uses this model in its Complex Behavior Support Program. According to program materials, admissions last from 12 to 36 weeks or more. In the first few weeks, staff conduct a functional analysis to determine why a child exhibits certain behaviors. Caregiver training usually begins halfway through the stay. Families also have responsibilities, including a home visit in the first week, weekly meetings with caregivers, and follow-up appointments that can continue until the patient turns 22.

A published comparison is available from Kennedy Krieger Institute’s neurobehavioral continuum. Its 16-bed inpatient unit, where about 75% of patients have an autism diagnosis, accepts referrals from across the country. A 2022 paper describes patients who had “exhausted all local outpatient and inpatient services in their state.” Kennedy Krieger has reported that about 88% of inpatients achieved at least an 80% reduction in their main target behaviors.

Why Payers Are Watching

The timing of these findings is important. The HHS Office of Inspector General (OIG) has completed four state audits of Medicaid ABA payments since December 2024, uncovering at least $56 million in improper payments in Indiana, $18.5 million in Wisconsin, $45.6 million in Maine, and $77.8 million in Colorado, for a total of nearly $198 million. In all four states, every one of 100 sampled enrollee-months, meaning one child’s claims in a single month, had at least one improper or possibly improper claim.

Some states have taken action on costs. In Indiana, Medicaid ABA spending exceeded $35 million in January 2026 alone for more than 6,000 enrollees. The state froze enrollment of new ABA provider agencies for at least six months starting June 6, 2026, a move approved by CMS, according to the Indiana Capital Chronicle. Eric Miller, deputy secretary of Indiana’s Family and Social Services Administration (FSSA), said the state “has seen an incredible surge in ABA spending” and called the freeze a “targeted, responsible step.”

UnitedHealth’s Optum unit has also taken steps. Internal Optum documents reported by ProPublica in December 2024 described plans to limit access to ABA. In some states, the company is moving to cut more than 40% of in-network provider groups from Medicaid networks and created an “approach to authorizing fewer units than requested,” even though the same documents called ABA the “evidence-based gold standard treatment” for children who need it.

These debates focus on documentation, cost, and the amount of care, not on whether treatment actually changes behavior. The new study addresses a question the audits do not: what happens to the patient. It reports outcomes for children whose aggression or self-injury puts them at immediate risk of harm. The paper does not say whether a 40-patient retrospective from one program will influence rate hearings or coverage disputes; the authors only describe the evidence as provisional.

The OIG’s ABA audit series still has four state reports in progress, with the series expected to finish in fiscal year 2027. Indiana’s enrollment freeze will last until early December 2026.

AT A GLANCE

SOURCES & REFERENCES