Correlations Between ADOS-2: Module 4 Comparison Severity Scores and Standardized Assessment Measures: A Preliminary Examination in Adults with Autism Spectrum Disorder

Background:  The recent update to the Autism Diagnostic Observation Schedule(ADOS-2) included a comparison severity score (CSS) to use as a quantifiable measure of ASD severity independent of one’s age and language functioning (Hus & Lord, 2014; Lord et al., 2012) to standardize the ADOS-2 total algorithm score for use as a dimensional measure of change. A recent article by Hus and Lord (2014) provided a revised algorithm and CSS scores for adults with ASD using Module 4. Comparison severity scores are reported in social affect (SA); restricted, repetitive behaviors (RRB); and Total ADOS-2 algorithm (Total) scores. This study examined the correlation between the CSS scores and individual characteristics of adults with ASD, as measured by standardized assessments of social responsivity reported by self and others, IQ, and memory. All measures were completed as part of an eligibility assessment for the Conte Center for Oxytocin and Social Cognition at Emory University.

Objectives:  To examine the correlation between standardized measures and ADOS-2: Module 4 CSS scores in order to determine which measure might be the best predictor of CSS scores.

Methods: Twenty-two male adults with ASD participated (M age = 27 years, 1 month, SD = 4 years, 1 month; M WASI-II Full Scale = 115.48, SD = 39.79). Standardized measures included the SRS-2 Social Communication and Interaction (SCI), Restricted Interests and Repetitive Behavior (RRB), and the Total SRS-2 t-scores from the self and other reports; standard scores for WAIS-II Verbal Comprehension, Perceptual Reasoning, and Full Scale-4. WMS-IV scaled sores on the Logical Memory Immediate, Logical Memory Delayed, Visual Reproduction Immediate and Visual Reproduction Delayed subtests. CSS scores were determined from Hus and Lord (2104) article. To facilitate comparison across assessments, all standard scores with the exception of ADOS-2 CSS scores were converted to z-scores by using the formula (score-mean)/standard deviation.

Results: Bivariate correlations indicated that only IQ scores predicted the CSS scores. In particular, we found a negative correlation between both sub-domains of IQ (Verbal Comprehension and Perceptual Reasoning) and the SA CSS (r = -.642, p < .01; r = -.531, p < .01, respectively). Negative correlations were also found for IQ sub-domains and the Total CSS (r = -.513, p < .01; r = -.442, p< .05, respectively). All other measures did not correlate with the CSS.

Conclusions: Our preliminary and exploratory results indicate that the CSS score of the ADOS-2: Module 4 is associated with intellectual abilities in adults with ASD. Our findings suggest that CSS or ADOS-2 scores are not necessarily reflecting only social functioning, but are also related to an individual’s intellectual characteristics. Further replications are needed on a larger sample. To optimize use of the ADOS-2 CSS scores as measures of change, researchers may need to control for intellectual ability in psychopharmacological and behavioral intervention studies that target core ASD symptoms.