A Comparison of Variability in Change Across Commonly Used Measures of Treatment Outcomes for Youth with ASD

Background:

In contrast to the several well-validated diagnostic and assessment measures of child symptomatology and functioning, there is little consensus regarding suitable measures to assess treatment outcomes (Matson, 2007). Within the treatment literature, measures of intelligence and adaptive skills measures are commonly used as indicators of change. However, their use present several issues, including questions regarding the sensitivity to changes (McConachie et al., 2015), especially given that many of these measures were designed to measure relatively stable constructs over time (e.g., intelligence). Given the significant empirical emphasis on evaluating intervention programs for youth with ASD, the identification of sensitive, suitable measures of treatment outcomes is critical.

Objectives:

The aim of the present study is to examine variability in rate of change observed across various outcome measures in a randomized trial of classroom-based ASD intervention, including standardized measures of cognitive and adaptive abilities.

Methods:

Participants include 263 students enrolled in a large effectiveness trial of a naturalistic behavioral intervention in school settings. Students were 3-11 year old (M = 5.9 SD = 2.24) and being served under the autism educational category in public school programs. Treatment outcome measures included: 1) standardized cognitive measures (either the Differential Abilities Scale-II or the Mullen Scales of Early Learning, as appropriate for developmental level) and the Vineland Adaptive Behavior Scales-II (VABS), as completed by the student’s teacher. Each measure was administered in the fall and spring of the school year. Progress across the school year on individual student goals selected by the teacher and assessed via Goal Attainment Scaling (GAS goals) were also included as a measure of change.

Results:

Results indicate the largest change in scores for the cognitive measures as compared to the VABS-II. Examination of associations between corresponding subscales from the cognitive measures and the VABS-II indicate a significant relationship between rates of change in receptive language but not expressive language. Further, whereas a positive association was observed between receptive language change and VABS-II expressive language, there was a negative association between expressive language and VABS-II receptive language. In terms of changes in academic functioning, there was a strong positive association between receptive language and improved academic skills; this is in contrast to the negative association observed with the VABS-II language scales. Further associations as well as potential moderating variables will be discussed.

Conclusions:

The current results suggest variability in rates of change across various common measures of treatment outcomes as data demonstrated by an inconsistent pattern of association between specific cognitive and adaptive abilities. Overall, these results lend support to the sensitivity to change in these commonly used treatment outcome measures. However, the inconsistencies seen in rate of change or progress across measures warrants further exploration.