In July, JAMA Pediatrics published an article from Dr. Michael Sandbank (University of North Carolina at Chapel Hill) and other researchers titled "Determining Associations Between Intervention Amount and Outcomes for Young Autistic Children: A Meta-Analysis."

The analysis by Sandbank et. al concluded there was insignificant evidence that children with autism will benefit from intensive intervention. However, three eminent researchers have questioned Sandbank's conclusions, highlighting flaws in the analysis.

Dr. Thomas Frazier (John Carroll University), Dr. L cey C. Chetcuti (Stanford University), and Dr. Mirko Uljarevic (Stanford University) published a response titled "Evidence that Intervention Dosage Does Associate with Better Outcomes in Autism."

According to Frazier et. al, "the primary study conclusion is heavily influenced by a significant selective sampling problem, which was not properly evaluated."

They add that Sandbank et. al failed "to fully consider one of the most important child characteristics, the child's capacity for processing and learning new information in their meta-regression models."

A re-analysis by Frazier et. al "shows that studies in the lower tercile for each intervention dosage metric had higher IQ scores than studies in the upper tercile. This is a major problem because pre-intervention IQ (cognitive ability) has previously been shown to be a major predictor of intervention outcome.

"Inclusion of IQ in the re-analysis of Sandbank data shows consistently significant, positive, and clinically-meaningful associations between intervention quantity and outcomes when IQ is included in the models."

Frazier et. al conclude, therefore, that "intervention dosage evidently does matter and must be factored into the design, evaluation, and implementation of treatment programs for individuals with autism spectrum disorder."