Challenges in Evaluating Improvements in Soars-B: Study of Oxytocin in ASD for Enhancing Reciprocal Social Behaviors

Background: The NIH-funded SOARS-B study was designed to evaluate the safety of six-month, twice-daily oxytocin across the full spectrum of children with autism spectrum disorders and to inform design of a future pivotal efficacy study with regard to inclusion criteria and primary outcome. The study included 277 children: 132 were minimally verbal and ages were evenly distributed between 3 and 17 years. In 2011, as now, there was no consensus about the most appropriate way to assess long-term changes in reciprocal social behaviors. The choice of a primary outcome measure was further complicated by the diversity of the planned sample, use of multiple sites and finite financial resources available for the project

Objectives: We sought a single primary outcome measure that could be reliably assessed across the pediatric age range and the full spectrum of cognitive and communicative abilities. The ideal outcome measure would also have the ability to be repeated multiple times over 24 weeks without learning effects and be sensitive to clinically meaningful changes in reciprocal social behaviors. It would measure positive behaviors as well as problem behaviors related to less well developed social skills interests and abilities.

Methods: Available literature was reviewed to determine the performance of current social behavior measures in prior ASD treatment studies across the age and cognitive range planned for SOARS-B. The National Database for Autism Research (NDAR) was used to study the variability of measures in a broader population of children with ASD. We chose a modified 13 item version of the Aberrant Behavior Checklist (ABC) lethargy subscale (removing the 3 items that were associated with low physical activity and referred to as the ABC-SW) as the primary outcome. Multiple other social measures were also used to assess social behaviors. Descriptive statistics for changes in different social measures over time during the trial and correlations between the changes in these different measures will be provided for the entire sample and key subsamples (e.g. minimally verbal). The changes in these measures also will be correlated with the Clinical Global Impressions Scale Improvement item (CGI-I) and a continuous visual analog scale of change in social behavior to help define the best measure for future studies.

Results: The analyses of the study results are currently in progress. We will report on the characteristics of our primary outcome measure - change in the ABC-SW - during the trial, its relationship with changes in other measures of social functioning including the Social Reciprocity Scale (SRS), the Pervasive Developmental Disorders Behavior Index Screening Version, the Vineland Adaptive Behavior Scales, and the CGI-I.

Conclusions: Choosing a single well-validated measure to assess change in social behavior across a broad range of functional and developmental levels is challenging. Most available measures rely on caregiver report and cover a limited range of social behaviors. More objective social outcomes such as eye gaze tracking, EEG, or behavioral tasks may be limited by stimuli appropriateness across development, multi-site standardization, and cost.