This commentary addresses three issues raised by a recently published multisite clinical trial designed to test replicability of a comprehensive intervention, the Early Start Denver Model, designed for very young children with ASD.

First, intervention targets and measures may focus on the behavioral symptoms of a condition or on downstream effects of the condition. This study focused on downstream effects on children’s development in those areas most affected by ASD: language and IQ, because of the relationships between standardized scores of language and IQ in childhood on outcomes later in life.

The second issue involves expected effects of replication methods and procedures. Rigorous replication studies tend not to result in the effect sizes reported in an initial report of a new behavioral treatment because of effects on the data related to increased number of subjects, added rigor of measures, increased rigor of analysis, and increased protections to prevent sources of bias. This study followed this pattern, in that it enrolled twice as many subjects as the initial study, had delivered treatment and gathered data from three sites rather than one, used an independent data team to analyze all the data, and used an intent to treat design that used data from all participants initially assigned to the two groups rather than only those who actually participated in the interventions over the full two years. The reduction in effect of treatment is seen in that this multisite study replicated the gain in language development in the experimental group found in the initial study but did not replicate other areas of significant advantage to the experimental group. Additionally, one site did not show group differences.

Third, the nature of comparison groups and treatment conditions in the present study differed in substantial ways from the initial study, seen most clearly in the increased amount of intervention, both public and private, that the comparison group in their communities. The decade of time between the initiation of the two studies was a period of great attention to ASD and appears to have had considerable positive effects on availability and quality of intervention services. The data from the present study compared to the initial study demonstrated large increases in amount of intervention and in developmental gains seen in the comparison group. In contrast, the data from the experimental groups in the two studies show very similar patterns of response to the experimental intervention on multiple developmental measures.

In summary, this large replication study used a rigorous multi-site design to test efficacy of intensive early intervention for toddlers just diagnosed with ASD. The study partially replicated the amount of change from age 2 to age 4 in language development, one of the best predictors of long-term outcomes, in young children with ASD who received a high quality, intensive, evidence-based intervention approach, the Early Start Denver Model, during their toddler and preschool years.