Characteristics of Autism Spectrum Disorder Across Time: Comparing Cohorts of Birth Date and Time of Assessment

Friday, May 13, 2016: 5:30 PM-7:00 PM
Hall A (Baltimore Convention Center)
T. White, T. E. Regan, K. Williams and M. R. Klinger, University of North Carolina at Chapel Hill, Chapel Hill, NC
Background: Over the course of this century, the diagnosis of autism spectrum disorder (ASD) has drastically increased.  In 2000, the rate of diagnosis was 1 in 150, and recent reports indicate the prevalence as 1 in 68 (Center for Disease Control, 2014).  In addition to growing numbers in the population, the autism profile has changed.  There has been an increase in the number of individuals diagnosed with normal to high intellectual functioning and adaptive behavior functioning (Freeman, Homme, Guthrie, & Zhang, 2009).  There is a need to evaluate these changes across the ASD population over time.

Objectives: This study examined associations of autism diagnostic characteristics with date of birth and date of evaluation within a large clinical sample.  Three diagnostic characteristics were included in analysis: autism symptom severity, adaptive functioning, and cognitive functioning.  The goals were to (1) determine if changes in diagnostic profile appearance were tied to date of birth, indicating a birth cohort effect, and (2) determine if changes in diagnostic profile appearance were tied to date of evaluation, indicating changing diagnostic criteria across time.

Methods: This study used secondary data analysis of assessment data collected from the statewide UNC TEACCH Autism Program registry. A records review was conducted to compare profiles of individuals evaluated in clinics.  The sample includes evaluations starting in 2000 through 2015.  Data from nearly 3000 individuals were used in this study.  A preliminary hierarchical multiple regression analysis was performed in order to investigate the ability to predict three facets of the autism profile (symptom severity, cognitive functioning, and adaptive functioning) based on date of diagnosis compared to date of birth, after controlling for both parent’s date of birth and the child’s gender.

Results: Data were analyzed separately for symptom severity (ADOS, N=1746 and CARS, N=1333), cognitive functioning (IQ, N=2929), and adaptive functioning (Vineland, N=2790). Hierarchical regression analyses were run for each variable with covariates (Parent 1 Date of Birth, Parent 2 Date of Birth, and Gender) entered first followed by predictor variables (Date of Evaluation and Child Date of Birth). Regression analyses found that both predictor variables contributed significant variance in the model for all four variables (ADOS: Child Date of Birth: beta = 0.315;  p < .001; Evaluation Date  beta= 0.265; p<.001) (CARS: CDoB beta= 0.237; p<.001; ED beta=0.117; p<.001) IQ: CDoB beta = .086; p < .001; ED beta = .357; p<.001) (Vineland: CDoB beta = .086; p < .001;  ED beta = .357; p<.001).

Conclusions: Preliminary results indicate a relatively consistent pattern for both date of evaluation and date of birth cohorts influencing the characteristics of individuals seen in a large network of community clinics.   Across the three domains (symptom severity, cognitive functioning, and adaptive behavior) all showed higher functioning for individuals diagnosed more recently and more adults. Part of this change may be due to changes in criteria (as indicated by evaluation date effects) and changes across birth cohorts (as indicated by child date of birth effects).