The Strongest Predictor of Outcome Was Hours. Regulators Are Calling Hours a Red Flag.
A new individual-participant meta-analysis finds substantial EIBI effects — and that intensity drove them. It lands the same month CMS flags excessive service hours.
Published in Autism Research in January 2026, Eldevik et al. conducted an individual participant data meta-analysis of Early Intensive Behavioral Intervention for children with autism aged 2–6 who received at least 12 months of treatment.
What they found
From 17 identified studies they obtained participant-level data from 15: 341 children who received EIBI and 280 in comparison groups.
- Adaptive behavior: effect size 0.66
- Intellectual functioning: effect size 0.87
- Autism severity (reduction): effect size 1.36
A significantly higher proportion of the EIBI group met criteria for statistically reliable change and scored in the non-clinical range after intervention, with a number needed to treat between 4.1 and 6.9. The authors propose benchmarks for evaluating interventions on this basis.
The caveat the authors state plainly
Every included study carried a serious risk of bias due to lack of random assignment. The authors say so directly, and say uncertainty remains because of it. Anyone citing these effect sizes should cite that sentence too.
The tension worth naming
The finding the authors highlight is that treatment intensity significantly contributed to changes across all outcome measures — weekly hours were the strongest group-level predictor of outcome.
Seven months later, CMS handed states a toolkit naming excessive service hours as a documentation red flag. Both things can be true: hours drive outcomes, and hours are also where improper billing hides. But practices are going to be asked to justify intensity to payers at exactly the moment the evidence for intensity got stronger.
What you must know or do
The practical response is not to argue the red flag. It is to make sure the file answers it before anyone asks.
For every client you bill at high intensity, two things should already be in the record: progress data showing movement against a baseline across the authorization period, and a clinical rationale for those hours written at the time they were set. A rationale composed for an auditor reads exactly like one composed for an auditor.
The evidence that intensity drives outcomes is stronger than it was a year ago. That helps you only where your own caseload can demonstrate it — the meta-analysis is an argument about children in general, and a payer is asking about one child in particular.