Research Review: Outcomes of 24- to 36-month-old children with autism spectrum disorder vary by ascertainment strategy: a systematic review and meta-analysis.

Abstract

Despite widespread recommendations for early surveillance of risk for autism spectrum disorder (ASD), no research to date has shown that early surveillance leads to better clinical outcomes. Preliminary research has suggested that children with ASD ascertained via prospective follow-up have better outcomes than those ascertained via community referral. Because prospective studies include early surveillance, by comparing outcomes of children with ASD across ascertainment strategies, we may gain insight into the effects of early surveillance relative to its absence. A systematic review was conducted to identify studies reporting outcomes of 24- to 36-month-olds with ASD ascertained via prospective follow-up, community referral, or universal screening. A meta-analysis using a random effects model was used to calculate overall effect size estimates for developmental level and symptom severity across ascertainment cohorts. Eleven prospective, ten community referral, and eight universal screening studies were identified, reporting on 1,658 toddlers with ASD. We found no differences in outcomes between community referral and universal screening studies. Relative to both, prospective studies reported significantly higher developmental levels and lower symptom severities. Outcomes of young children with ASD ascertained via prospective follow-up are better than those of children with ASD recruited via community referral or universal screening. Although we discuss why sampling bias is not likely the driving force behind these findings, we cannot rule out the possibility that sampling bias contributes to the observed differences; future studies should probe the effects of sociodemographic variables on clinical outcomes as a function of ascertainment strategy. This limitation notwithstanding, our results raise the possibility that prospective follow-up may confer a 'surveillance effect' that contributes to improved developmental and diagnostic outcomes in children with ASD. Future research should test this hypothesis and determine the specific mechanism by which surveillance may improve outcomes.