Testability in Young Autistic Children: A Descriptive Longitudinal Study Comparing Conventional and Strength-Based Assessment.

Background: DSM-5 autism spectrum (AS) disorder diagnosis is often done early in childhood and must specify whether there is an accompanying intellectual disability (APA, 2013). However, conventional intellectual assessment at such an early age bears many challenges and may not accurately represent the potential of young autistic children, whose developmental paths are highly atypical. For example, many autistic children are considered as having intellectual delay at preschool age when assessed with conventional tests. However, many of them will no longer exhibit an intellectual delay or disability at school age.

Objectives: We first wanted to describe testability at the age of diagnosis (T0) and a year later (T1) in a group of young AS and typically developing (TD) children using WPPSI-IV as conventional assessment and Raven Coloured Progressive Matrices (RCPM) as strength-based assessment. Second, we wanted to determine the proportion of AS and TD children with scores in the intellectual delay range (QI<70 or percentile rank<=2%) when assessed with WPPSI-IV compared with RCPM.

Methods: At T0, 100 children (50AS; 50 NT) aged 30-70 months (mean 54 months) were tested with both WPPSI-IV and RCPM. Data collection for T1(one year after T0) is still ongoing, but 46 children (20AS; 26NT) have already completed T1 with the same two tests. Of these, 38 children (12AS; 26 NT) were able to complete all tests at both time points. Both groups are matched on age at both time points (ps>.05).

Results: At T0, all 50 TD children who attempted to complete WPPSI-IV and RCPM were able to complete both tests. The 26 TD children who were called back a year later were also all able to complete both tests. Out of the 50 AS children who were tested at baseline (T0), 58% (29/50) were able to complete WPPSI-IV whereas 66% (33/50) were able to complete RCPM. A year later (T1) testability of AS increases for both tests as 75% (15/20) were able to complete WPPSI-IV and RCPM. Also, the proportion of AS children with scores in the intellectual delay range (IQ<70 or percentile rank<=2%) is higher when using WPPSI-IV [14% (7/50)] compared to RCPM [2% (1/50)] at T0. A year later, this percentage decreases to 10% (2/20) using WPPSI-IV and to 0% (0/20) using RCPM in the AS group. Linear regression analyses are planned to determine whether performance on WPPSI-IV and on RCPM could predict the evolution of cognitive potential, operationalised as change between WPPSI-IV standard score at T0 and T1.

Conclusions: In the AS group, the performance and ability to complete WPPSI-IV increases a year after baseline, as children get older. Also, administration conditions specific to RCPM seem to favour AS children in their comprehension and compliance to the test relative to WPPSI-IV. Thus, a strength-based approach could help bypass some challenges of assessing young AS children at the age of diagnosis. Our results also suggest that assessing AS children at such an early age is difficult and that waiting until school age to assess their cognitive potential could help increase their testability.