Six Meta-Analyses, 2,241 Children: Early ABA Lifted IQ and Adaptive Behavior

A JADD umbrella review by Reichow, Barton and Song found consistent effects on the two primary outcomes. On communication, social skills, daily living and symptoms, the evidence was thinner and more variable.

On October 1 the Journal of Autism and Developmental Disorders published an umbrella review by Brian Reichow, Erin Barton and Jinwei Song of every systematic review with a meta-analysis of comprehensive ABA for young autistic children they could find. That came to six reviews, 36 unique studies and 2,241 children, with searches running through December 2025.

The turn: strong where payers ask least, thinner where they ask most

The timing matters. The CMS toolkit asks states to tie ABA to measurable outcomes, and payers like Molina now want progress attributable to ABA, “not age alone.” This review gives a firm answer on IQ and adaptive behavior, in studies with comparison groups. It gives a much weaker one on communication, social skills and symptom severity, which is where a lot of treatment-plan goals and payer criteria sit.

What it found

  • IQ (primary outcome): three meta-analyses reported standardized mean differences above 0.50. Three reported mean differences above 11.98 points on a scale with an SD of 15.
  • Adaptive behavior (primary outcome): three reported standardized mean differences above 0.35. Three reported mean differences of 5.92 points or more.
  • Communication, social, daily living and symptoms (secondary outcomes): “reported less frequently across reviews with greater variability in findings.”
  • The authors conclude that children receiving the intervention did better than comparison children on IQ and adaptive behavior “for a majority of the participants who received the intervention.”
  • The protocol was registered in advance with PROSPERO, and the authors assessed risk of bias.

The limits

The authors set two limits themselves. The secondary outcomes are sparse and inconsistent, and the benefit holds for a majority of children, not all of them. The six reviews also overlap heavily: the corrected covered area is 20.56%, which is conventionally read as very high overlap. So six reviews agreeing is largely the same body of studies being pooled several times, not six independent replications. The abstract reports no analysis of intensity or hours, so this doesn't settle the dose question raised by the two pooled analyses we covered in August. The full text is behind a paywall, and the review-by-review risk-of-bias ratings aren't in the abstract.

What you must know or do

  • Owners and clinical directors answering a payer or state challenge to the evidence base: cite this review (doi:10.1007/s10803-026-07551-6) for IQ and adaptive behavior in young children receiving comprehensive programs. Don't cite it for social or communication goals, or for older children or focused programs, because it doesn't cover them.
  • BCBAs: pull your last five reassessment reports for young clients in comprehensive programs. Check whether each one reports a standardized adaptive behavior score. That's the outcome with the strongest evidence here, and the measure a reviewer is most likely to recognize.
  • Anyone writing intake or parent materials: if they quote an IQ gain, make sure it's framed as a group average. The authors' own wording is “a majority” of participants, not every child.