Early Intervention Project: can its claims be substantiated and its effects replicated?

Wilkinson et al. (1998) took a hard look at the famous Early Intervention Project. The project had claimed that 47 % of toddlers with autism "recovered" after intense behavioral treatment. The authors checked if those bold claims held up under close inspection. They read every report on the project and graded the research methods.

The review found big holes in the EIP research. There was no true control group. The diagnosis methods kept changing. Outcome tests were not the same for all kids. In short, the study could not prove that the intervention caused the gains. The "recovery" claim rested on shaky ground.

Lord et al. (2005) later made the same plea: autism intervention studies must use randomized controlled trials. Their paper echoes the 1998 warning seven years down the road.

Méla et al. (2019) did what the target paper asked for. They tracked kids who got low-dose EIBI and found IQ gains faded while autism symptoms returned. This real-world data supports the 1998 call for tougher proof before promising recovery.

Chung et al. (2024) still lists only EIBI and tailored ABA as "best-supported" twenty-six years later. The field has not moved far beyond the cautions first voiced in 1998.

When you see a new autism intervention boasting miracle rates, pause. Ask for a peer-reviewed RCT with clear diagnostic criteria and masked outcome assessors. Until that bar is met, stick to well-replicated models like EIBI or parent-mediated ABA that have survived this kind of scrutiny. Your clients deserve nothing less.