Is There an Optimal Developmental Path in Autism?

Background:   Two striking features of autism are diversity in reported developmental paths and diversity in reported autistic outcomes, from outstanding to extremely poor. Current approaches to autism posit the existence of an optimal developmental path, against which the potential of autistics can be judged and thus their outcomes predicted. However, long-term follow-up studies of diagnosed autistics present a more complex picture (e.g., Howlin, 2011), while existing epidemiology suggests that many older autistics are undiagnosed (e.g., Brugha et al., 2011; Kim et al., 2011). Further, the apparent consensus that early development, particularly speech onset timing, cannot be used for diagnostic subgrouping (APA, 2012) calls into question its ostensibly crucial role in determining autistic outcomes.

Objectives:   We aimed to verify whether speech onset timing is related to later (school-aged and older) outcomes of concern in autistic spectrum individuals subgrouped according to presence or absence of speech development anomalies. We also aimed to compare later outcomes as assessed via three different instruments, in autistic children and autistic adults subgrouped in the same way. This study builds on data we presented at IMFAR 2007.

Methods:   We retrieved three data sets from consecutive cases from the Riviere-des-Prairies Hospital database who were six years of age or older and had an autism (with speech development anomalies) or Asperger syndrome (without speech development anomalies; normal-range Wechsler IQ) best estimate clinical diagnosis and ADI-R evaluation above threshold for autism. First and second sets included all individuals with valid age of first words and phrases ADI-R data, as well as Wechsler IQ (first set) or Vineland scores (second set). The third set included all individuals with Wechsler IQ, Raven’s Progressive Matrices, and Vineland adaptive behavior scores.

Results:   For autistics (N=90) there was no significant correlation between age at first words or phrases and Wechsler full-scale, verbal, and performance IQ, or (N=61) Vineland scores. There was no significant difference between autistics with 70 or above versus below 70 FSIQ, in age of first words (mean respectively 29.3 versus 32.8 months, p=.46) or phrases (41.7 versus 43.6 months, p=.64). For Asperger individuals (N=39), there was no significant correlation between age of first words and FSIQ, VIQ or PIQ, or between age of first phrases and VIQ. However, there was a significant correlation between age of first phrases and both FSIQ and PIQ. As with autistics, for Vineland scores (N=29) there was no correlation with either age at first words or phrases. Comparisons of Wechsler, Vineland, and Raven scores in autistic and Asperger participants revealed dramatic discrepancies across measures in both adults and children, and higher Wechsler and Raven scores in autistic adults versus children.

Conclusions:   Speech onset timing is unrelated to later IQ or adaptive behavior scores in autistic individuals defined by speech development anomalies. Different instruments can give dramatically different portraits of autism spectrum outcomes. Data from short-term studies and/or biased samples (e.g., samples excluding later-diagnosed autistics) may be distorting current conceptions of outcome predictors and limiting our knowledge of how atypical developmental paths lead to outcomes in autism.