Longitudinal Change in Symptom Domains by ASD Severity Trajectories

Background: The Autism Diagnostic Observation Schedule (ADOS) is often a primary phenotype measure in research samples, though the modular format of this measure makes longitudinal data comparison difficult.

Objectives: To plot longitudinal trajectories of ASD severity among children and adolescents using standardized ADOS scores, then to examine raw total changes over time in ADOS Social Affect (SA) and Restricted Repetitive Behavior (RRB) domains within each trajectory class.  

Methods: The standardized ADOS severity metric reported by Gotham, Pickles, & Lord (in press) was applied to 1026 cases of data collected longitudinally from 345 individuals referred for ASD. Standardized scores were fitted for latent classes of severity trajectories with and without covariates. Within each latent class, trends in Social Affect and Restricted, Repetitive Behavior raw totals as a function of age were described in order to examine contribution of each domain to overall severity patterns over time.

Results: A four class model of ASD severity trajectories best represented the observed data, including a persistent high severity class, a moderately severe class, and two classes that respectively increased or decreased in ASD severity over time. In the persistently high and moderately severe classes, average RRB scores remain approximately stable while SA scores appear to decrease slightly over time. Within the increasing and decreasing severity classes, both domains appear to increase or decrease respectively.

Conclusions:  If replicated, identified classes of autism severity trajectory may help in making clinical prognoses and subtyping samples for neurobiological and genetic research. Insight into the direction, magnitude, and age periods associated with changes in ASD domain severity may be able to guide intervention efforts, as well as the study of developmental trajectory of these disorders.