Predictors of Functioning in Preschool and School Age Children with ASD

Background:  

Many longitudinal studies thus have examined the stability of the ASD diagnosis over time, yet occurred at a 1 to 2 year period, at preschool age, when the symptomology is most prototypical (Thurm, Lord, Lee & Newschaffer 2006). There is a need for longitudinal research on the trajectories of development and predictors of functioning for school age children that explore the core developmental factors involved. Smaller studies have looked at measured outcome for three time points (Charmen, 2005); however, most have focused on outcome rather than level of functioning across time.

Objectives:  

To examine what degree measures of early communication, language, cognitive and social competence predict later outcomes in nonverbal functioning and autism severity in a cohort of children with ASD.

Methods:  

Participants included 74 children with ASD who were initially derived from a sample of children who were evaluated at ages two (toddler), and four (preschool), as part of participation in the NAAR/CPEA and STAART funded Yale projects.  These participants were seen again at age eight (school age) as part of their participation in the NICHD funded study on the Developmental Processes, Trajectories, and Predictors of Outcome in a Longitudinal Cohort.  The assessments at ages two and four included measures of autism, adaptive, cognitive and language functioning and were used to examine the predictability across time of autism severity and nonverbal functioningfor this cohort.

Results:  

Pearson’s correlations were conducted to assess the cross-time association between a variety measures. All characterization measures demonstrated strong correlations from 1^st to 2^nd visits and from 2^nd to 3^rdvisits.  Multiple linear stepwise regressions were further run to identify significant relationships over time among autism severity and cognitive development with measures of language, social, adaptive and cognitive functioning.

Preliminary results show a shift in predictors based on point of assessment. 30% of the variance in ADOS calibrated severity score (CSS) measured at age four  was predicted by the CSS at age two (R²=.296, F(1,67)=28.11, p<.000) whereas  a shift to 38% of the variance in calibrated severity scores at age eight  was predicted by socialization and functional communication (Vineland)  in combination with CSS at age four (R²=.381, F(3,57)=11.68, p<.000).

47% of the variance in nonverbal performance at age four was predicted by, non-verbal developmental quotient (Mullen),  functional communication (Vineland), and autism severity (CSS) at age two (R²=.473, F(3,63)=18.82, p<.000). By age eight, however,  VDQ (Mullen)  is most strongly predictive of nonverbal performance with 57% of the variance. (R²=.570, F(1,64)=84.82, p<.000).

Conclusions:  

The combinations of multiple source of information including direct assessment and parent interview lead to more stable prediction of functioning, particularly in the transition from preschool to school age. The identification of early mediators will also inform research on the predictive power of various measures and may contribute to establishing goals for early intervention services.