Background. Individuals with Autism Spectrum Disorders (ASD) show deficits in adaptive functioning, particularly in the domains of socialization and communication, but also more generally. Assessments of adaptive behavior are therefore used routinely in clinical practice to quantify the severity of ASD symptoms and to track changes in functioning over time. Interestingly, despite the clinical importance of this measure, there are few longitudinal studies of adaptive behavior in children with ASD. Studies that have been published show that standardized adaptive behavior scores tend not to improve or to decline over time despite standard therapeutic intervention, indicating widening discrepancies between age-expected and observed adaptive skills.

Objective. We sought to characterize longitudinal changes in adaptive behavior in a clinical cohort of children with diagnoses of ASD followed at Cortica centers. These children received Applied Behavior Analysis (ABA) therapy as part of a comprehensive program that also included optimizing pharmacologic and behavioral management of associated medical conditions (such as epilepsy, gastrointestinal issues, and sleep disturbance) as well as other developmental therapies (including occupational therapy and speech-language therapy, as clinically indicated). We hypothesized that this comprehensive, personalized model would lead to improved outcomes compared to published longitudinal studies representing typical care.

Methods. We identified children within the Cortica population who (1) had diagnoses of ASD, (2) had at least two discrete assessments of adaptive behavior using the Vineland Adaptive Behavior Scales (VABS), and (3) were between 18 months and 10 years of age at the time of assessment. Children were classified as having Mild, Moderate, or Severe symptoms based on the Adaptive Behavior Composite (ABC) standard score at the first assessment. The same criteria were used to identify children included in a dataset from a longitudinal study of adaptive behavior archived in the National Database for Autism Research (NDAR)6 (Table 1). Because of the small number of children with Mild symptoms, only children with Moderate and Severe symptoms were included in analyses.

Table 1: Characteristics of children included in each cohort.

For each child, we then calculated differences between ABC standard scores at consecutive assessments (e.g. ABC2 – ABC1, … , ABCn+1 – ABCn). Each Difference Score was treated as a single observation (Table 2).

Table 2. Characteristics of observations included in analyses.

Finally, we constructed an analysis of variance (ANOVA) with ABC Difference Score as the dependent variable, and Cohort and Level of symptoms as predictor variables. To account for differences in interval between assessments within and across cohorts, this was included as a covariate.

Results. There was a significant main effect of Cohort (F = 4.42, p < 0.04) indicating greater improvement in ABC scores for the Cortica cohort (mean 1.4 ± 0.5) compared to NDAR (mean 0.1 ± 0.3) (Figure 1). There was no main effect of Level and no interaction between Level and Cohort, indicating that there was greater improvement in the Cortica cohort compared to NDAR for children with both Moderate and Severe symptoms.

Figure 1. Differences in ABC between Cortica and NDAR cohorts.

We subsequently conducted t-tests to assess for differences between cohorts in VABS subscales including Communication (COM), Socialization (SOC), and Daily Living Skills (DLS). This analysis confirmed significantly greater improvement for the Cortica cohort compared to NDAR in COM and SOC, but not DLS (Figure 2, Table 3).

Figure 2. Differences in VABS subscales between Cortica and NDAR cohorts.

Table 3. Differences in VABS subscales between Cortica and NDAR cohorts.

Discussion. We compared development of adaptive behavior between a cohort of children at Cortica and a reference cohort drawn from a longitudinal study of adaptive behavior archived in NDAR. In the NDAR cohort, there was little net change in ABC scores over time, consistent with the observation by that study’s authors that children with low initial scores tended to have decreasing scores over time, while those with initial scores in the moderately impaired range tended to be stable6. This mirrors findings of other recent studies of adaptive behavior development in children with ASD5.

By contrast, children in the Cortica cohort tended to show increasing ABC scores between consecutive assessments, regardless of level of symptom severity. Analysis of variance demonstrated that this pattern was significantly different from the NDAR cohort. Similar trends were present for individual subscales of the VABS including Communication and Socialization, though not for Daily Living Skills. (Of note, DLS Difference Scores showed greater variability in the NDAR cohort than any other measure, affecting the significance of a direct comparison.)

One reason for this difference may be the type of therapeutic intervention received by children in each cohort. Children in the Cortica cohort participated in an individualized and coordinated program of therapies (Table 4) under the clinical direction of a pediatric neurologist, who was also responsible for management of medications and associated medical conditions. We note that the intensity of behavior therapy was below practice guidelines for ABA, which recommend 10-25 hours per week for focused interventions and 30-40 hours of direct therapy per week for comprehensive interventions7. The NDAR dataset does not include information about therapies received, but it can be assumed that these children received care that is typical in the community, with fragmented services across multiple providers.

Table 4. Average hours of therapy per week for children in Cortica cohort with Moderate and Severe symptoms.

Conclusion. A comprehensive, personalized, and medically based care model for children with ASD results in significantly improved adaptive behavior development compared to a reference cohort of children receiving typical care.


  1. Volkmar F, Sparrow S, Goudreau D, Cicchetti DV, Paul R, Cohen D. Social deficits in autism: an operational approach using the Vineland Adaptive Behavior Scales. J Am Acad Child Adolesc Psychiatry1987;26:156-61.

  2. Kenworthy L, Case L, Harms MB, Martin A, Wallace GL. Adaptive behavior ratings correlate with symptomatology and IQ among individuals with high-functioning autism spectrum disorders. J Autism Dev Disord. 2010;40:416-23.

  3. Zander E, Bolte S. The new DSM-5 impairment criterion: a challenge to early autism spectrum disorder diagnosis? J Autism Dev Disord. 2015;45:3634-43.

  4. Scahill L. Diagnosis and evaluation of pervasive developmental disorders. J Clin Psychiatry 2005;66(Supp 10):19-25.

  5. Szatmari P, Georgiades S, Duku E, Bennett TA, Bryson S, Fombonne E et al. Developmental trajectories of symptom severity and adaptive functioning in an inception cohort of preschool children with autism spectrum disorder. JAMA Psychiatry 2015;72:276-83.

  6. Farmer C, Swineford L, Swedo SE, Thurm A. Classifying and characterizing the development of adaptive behavior in a naturalistic longitudinal study of young children with autism. J Neurodev Dis 2018;10:1. DOI 10.1186/s11689-017-9222-9

  7. Applied Behavior Analysis Treatment of Autism Spectrum Disorder: Practice Guidelines for Healthcare Funders and Managers, 2nd Edition. Behavior Analyst Certification Board, Inc. 2014