Population-Based Screening for Autism Spectrum Disorder Using the Social Communication Questionnaire

Background:  The Social Communication Questionnaire (Rutter, Bailey, and Lord, 2003) is a 40-item checklist designed to identify children who may be at risk of having autism spectrum disorder (ASD).

Objectives:  To describe the distribution of SCQ scores in a population of school aged children (8 years of age) from a racially and economically diverse area in South Carolina (SC).

Methods:  The South Carolina Children’s Educational Surveillance Study (SUCCESS) is an ongoing population-based study that will determine the prevalence of Autism Spectrum Disorder (ASD) among a population of children born in 2004 and living in one of three counties surrounding Charleston, SC. The region is racially and economically diverse (59% White, 32% Black, 7% Hispanic, and 2% other; 33% schools with Title 1 status). The SCQ was distributed to parents of eligible children in approximately half of the public and private schools in the catchment area (the study is in progress in the remaining schools). Ongoing efforts have been made to reach home-schooling families through community outreach and home-school associations. Rate of participation in the screening phase for the first three waves (out of six planned waves) of the study varied by school and ranged from 22%-100% (overall response rate thus far = 51%). Schools with low response rates (<50%) are currently being targeted with special interventions to improve responding. 

Results:  In the first three waves of the SUCCESS study (n = 1498), 2.2% of youth were at very high risk for ASD (SCQ>21); 5.41% were at risk (SCQ 15-21); and 20.83% fell just below the threshold for risk on the SCQ (SCQ 8-14). Males were more likely to score in the at-risk range than females (x²=6.87, p=0.009). Minority children had a higher average SCQ score than white children, and were more likely to be identified as being at-risk for ASD (x²=5.91, p=0.015). Hispanic children were also more likely to be identified as being at-risk compared to non-Hispanic children (x²=8.31, p=0.0039).  In addition, children attending Title 1 schools (those with a high percentage of low-income students) also had higher average SCQ scores and were more likely to be at risk for ASD than those children not attending a Title 1 school (x²=43.7, p<0.0001).  

Conclusions: Almost 8% of 8-year-old children in this general population sample were identified as being at-risk for ASD. Males, minority children, and children attending low-income schools had higher average SCQ scores, and were more likely to be identified as being at risk for ASD   Further evaluation is needed to determine if these increases accurately reflect higher ASD risk in these populations or if these differences can be accounted for by differential responding or measurement artifacts on the SCQ.  Future SUCCESS efforts will address these issues through the completion of  screening, and ongoing diagnostic evaluations of approximately 10% of the sample in order to establish the prevalence of ASD in the study area, compare DSM-IV and DSM-5 diagnoses, and determine the sensitivity and specificity of the SCQ as a screening instrument for ASD in population-based samples.