Objectives: This study assesses the predictors of diagnostic stability of a diagnosis of autistic disorder and social and communication adaptive behaviours at twelve months follow-up for preschool aged children enrolled in a comprehensive intervention trial. The trial was established to assess centre versus home-based intervention.
Methods: For the trial eligible (clinically diagnosed ASD, preschool age) children were either randomised to centre or home-based intervention or included as a wait list group. For this study outcome measures were diagnostic stability for children diagnosed with autistic disorder (N=55) and social and communication adaptive behaviour in all eligible children (N=84), measured by the Vineland Adaptive Behaviour Scales (VABS). All children were assessed at baseline and at twelve months follow-up using standardised measures. This study was funded by the Australian Research Council.
Results: Children had a mean age of 3.5 years (range 2.2–5.0, SD = 0.61) at baseline. Fifty-five children were diagnosed with autistic disorder at baseline using the Autism Diagnostic Observation Scale (ADOS). Of these children seven (13%) improved and were not diagnosed as autistic disorder using the ADOS at twelve months follow-up. Five were diagnosed as ASD and two as not ASD. Univariate analysis identified baseline Griffith Mental Developmental Scale (GMDS) developmental quotient, Pragmatics Profile (total Q range), and Reynell comprehension raw score as predictors of improvement, whereas VABS social and communication scores and the Reynell expression raw score were not statistically significant predictors. Following logistic regression analysis GMDS developmental quotient was the only significant predictor explaining approximately 20% of the improvement in diagnosis in this sample. Univariate predictors of VABS social and communication scores at twelve month follow-up were GMDS developmental quotient, and Pragmatics Profile, Reynell comprehension score, Reynell expression score and ADOS diagnostic classification but not treatment group. Using regression analysis only the GMDS developmental quotient and Pragmatics Profile persisted as significant predictors of both the VABS social and communication scores, with those two factors explaining 54% and 59% of the variation in follow-up scores respectively.
Conclusions: Developmental quotient as measured with the GMDS was an important predictor of diagnostic stability and adaptive behaviour for children with ASD in this study. Information about diagnostic features as well as a measure of overall developmental ability and pragmatic communication is needed to allow a meaningful conversation with parents about the likely short term outcome for their preschool aged child. Further distillation of other predictors of outcome and whether these predictors are the same for older children is also needed.