Evidence For Clinical Utility of a Dimensional Measure of Autistic Traits in Toddlers

Objective: The Video-Referenced Rating of Reciprocal Social Behavior (vrRSB), a parent-report metric of quantitative autistic traits (QATs) in toddlers, leverages an innovative video scoring achor to reliably measure heritable, continuous varriation in autistic traits within the general population. However, the vrRSB's ability to distinguish toddlers with and without autism spectrum disorder (ASD) has not been formally tested. Here we compared vrRSB scores in children with and without ASD in two samples: a community-ascertained research sample (RS) recruiting toddlers with and without ASD and a clinically referred sample (CS) consisting of toddlers with concerns for ASD. Methods: T-tests and logistic regressions indexed differences in vrRSB scores for toddlers with versus without ASD in both samples (RS: n ASD =92, n non-ASD =456; CS: n ASD =127, n non-ASD =44). Regression models additionally tested for differential functioning of the vrRSB relative to sex and race. ROC analysis compared classification performance of the vrRSB versus the Modified Checklist for Autism in Toddlers (M-CHAT). Results: Autistic toddlers exhibited significantly higher vrRSB scores in both samples: Cohen's d RS =3.68; Cohen's d CS =1.04, ps<.001. In regression models, vrRSB scores predicted ASD diagnosis (OR=2.59, p<.001) and scores did not function differentially by sex or race (ps>.293). The vrRSB exhibited similar area under the curve as the M-CHAT in both samples with higher sensitivity in the CS. Conclusion: Toddler QATs show predictive validity for autism diagnostic outcome. Consistent performance across sex and racial groups suggests promise for generalizable clinical interpretation and implementation. High sensitivity of the vrRSB supports utility for secondary screening in specialty evaluations.