Screening for Autism In Toddlers: A Follow-up Study of the EACH CHILD Cohort

Background:  

In the last decade there has been an emphasis on identifying the earliest signs of autism, which would improve early identification, lower the age at which children can receive early intervention, and possibly help identify etiologies.  To date, most of this research has been conducted on children either already diagnosed (e.g., retrospective research using early home videos of children with autism) or who were known to be at high risk (e.g., research on the infant siblings of children with autism).  Population-based studies of how ASD unfolds in the early months and years of life have not yet been published. Screening studies conducted in community settings can begin to shed light on this issue by examining screened children over the first few years of life.

Objectives:  

The present study follows children who were originally screened as toddlers through a large community-based pediatric practice using the MCHAT and CSBS-ITC.  Children were evaluated after their 3^rd birthday to examine diagnostic course and clinical presentation.

Methods:  

Of the 796 families screened through the original screening study (The EACH CHILD Study; under review), 51 families were eligible for an in-person screening based on questionnaire results; 30 chose to participate. All 51 families were re-contacted and asked to participate in a follow-up evaluation.  Twenty-four families participated in a three-hour diagnostic assessment including cognitive and adaptive testing and parent interviews.  Overall accuracy (Positive Predictive Value) of the study was calculated and group cognitive and adaptive profiles were examined.

Results:  

The PPV for the screening process as a whole was 0.90 for ASD (9 of 10 children were correctly identified as EAD).  PPV from questionnaire scores alone (i.e., positive screen scores on either the M-CHAT or ITC) would have been 0.06 (9/145), and PPV from the phone-follow-up would have been 0.43 (9/24).  The ASD group demonstrated lower scores on the nonverbal domains of the Mullen Scales of Early Learning and on the Socialization domain of the Vineland Adaptive Behavior Scales. 

Conclusions:  

Our results suggest that a systematic screening process in a large, community-based pediatric practice accurately identified toddlers who would go on to meet DSM-IV criteria for Autistic Disorder or PDDNOS after the age of 3 years, without over-identifying children or providing a significant number of unnecessary evaluations.  Larger systematic screening studies are needed to further examine general population samples for ASD.