The Fluid Nature of the Very Early ASD Phenotype: Examination of Diagnostic Stability of ASD within a General Population Cohort
Very early detection (between 12-24 months) and early treatment have been flagship goals of government regulatory bodies such as the IACC, parent, and advocacy groups alike. Determining how early autism can be accurately recognized, and the stability of ASD once it is suspected, is still unclear. Almost all that is known in this area comes from baby-sibling studies. One such study by Ozonoff and colleagues, noted strong diagnostic stability of toddlers identified as ASD at 18 months (92% stability), however approximately half of the children diagnosed with ASD at 3 years were not identified as ASD when previously tested at 18 or 24 months. The stability of diagnoses in such baby sibling studies are usually presented in dichotomous terms: ASD or Not ASD, making it challenging to understand how a toddler initially identified around 18 months as typical or with a non-ASD delay may end up with an ASD diagnosis at age 3 years or vice versa. Understanding not only the stability of ASD identification in toddlers, but how this designation may be blurred across diagnostic boundaries across the first 3 years of life in a general population cohort is essential.
(1) To examine diagnostic stability of ASD starting at 12 months in age in the general population. (2) To determine the proportion of changes across diagnostic boundaries.
545 toddlers at-risk for ASD and other delays between 12-24 months identified in the general population using the 1-Year Well-Baby Check-Up Approach, which centered around using the CSBS IT Checklist at pediatric well-baby visits, participated (M= 15mo). Toddlers participated in a comprehensive developmental evaluation and were followed every 6-9 months until their 3rd birthday when a final diagnosis was given (M= 37 mo). Developmental evaluations at every age were conducted by Ph.D. level licensed psychologists and diagnostic judgments for a range of delays (ASD, ASD features, language delay, global developmental delay, motor delay, etc.) at each age were tracked.
Overall, diagnostic stability of ASD between 12-24 months (~80%) was lower in this general population cohort in comparison to reports from baby sibling studies that report levels as high as 92%. Eight percent of toddlers initially considered ASD at intake age were given a designation of ASD-features at final diagnosis age, 8% showed a non-ASD delay (e.g., language delay), and 4% were considered typical. Similar to baby sibling studies, a large percentage of cases were not designated as ASD initially. Common diagnostic transitions included language delay to ASD (16% of final ASD sample) and global developmental delay to ASD (30%).
While diagnostic stability of ASD as identified in toddlers in the general population may not be as high as noted in baby sibling cohorts, all but 4% of cases retained some delay designation at final diagnostic age, suggesting that stability of having a delay in general is extremely high (>95%). The persistence of deficits in key skills such as language and social behavior of toddlers initially identified as ASD warrant early treatment regardless of final diagnostic designation.