The Social Communication Questionnaire (SCQ) As a Population-Based Screening Instrument for Autism Spectrum Disorder

Friday, May 13, 2016: 5:30 PM-7:00 PM
Hall A (Baltimore Convention Center)
C. C. Bradley1, A. D. Boan1, A. Wahlquist2, A. P. Cohen3, J. M. Charles1 and L. A. Carpenter4, (1)Pediatrics, Medical University of South Carolina, Charleston, SC, (2)MUSC, Charleston, SC, (3)University of Illinois at Urbana-Champaign, Champaign, IL, (4)Medical University of South Carolina, Charleston, SC
Background:  The Social Communication Questionnaire – Lifetime Version (SCQ) is a 40-item parent-report screening measure for Autism Spectrum Disorder (ASD) that is widely used in research and clinical practice.  Although the SCQ has been studied in both clinical and general population samples, the sample sizes have been small and there is limited data available regarding how demographic factors, such as gender, race, or SES might impact parent responses on the SCQ. 

Objectives:  (1) To determine the sensitivity and specificity of the SCQ as a screening instrument for ASD in a large, epidemiological sample of youth;  (2) To examine the impact of various demographic factors on SCQ scores.      

Methods:  Data for the present study comes from the South Carolina Children’s Educational Surveillance Study (SUCCESS) which was designed to assess the prevalence of ASD in children born in 2004 (aged 8-10 at the time of participation) through direct screening and evaluation.  SCQs were distributed to all children in the study area through schools.  Participants with SCQ≥15, as well as 20% of those with scores between 8 and 14 were invited to participate in a diagnostic assessment. 

Results:  Of the 8246 SCQ’s that were distributed, 4085 (50%) were returned. Total SCQ scores were available for 3720 participants (91%) with a mean of 6.09 (SD = 5.44, Range: 0 to 36) and a median score of 5.  Over 7% of the sample had an SCQ score of 15 or higher, with another 23% of SCQs falling between 8 and 14.  Participants were 49% male, 52% white non-Hispanic, 23% black non-Hispanic, 4% Hispanic, and 51% from Title 1 public schools (schools with a high percentage of low income students). Unsurprisingly, males had significantly higher mean SCQ scores than females (6.87 vs. 5.34, p<.0001).  Title 1 school status was also associated with higher SCQ scores (7.35 vs. 5.03, p<.0001).  Race and ethnicity data was available for less than half of the sample due to differences in data availability in each school district.  Non-Hispanic black families reported higher SCQ scores than non-Hispanic white families (7.65 vs. 5.08, p<.0001) and families who identified as Hispanic also had significantly higher SCQ scores (9.41 vs. 6.24, p<.0001).  For youth with SCQ≥15 who completed an in-person assessment, the unadjusted OR = 1.32, indicating a 32% increase in risk for meeting DSM-5 criteria for ASD for each additional SCQ point above 15. The mean SCQ score amongst youth diagnosed with DSM-5 ASD was 23.54.

Conclusions:  This study includes one of the largest samples ever to utilize the SCQ for population-based screening for ASD.  Significant mean score differences based on demographic factors including race, ethnicity, and Title 1 school status suggest that this instrument may perform differently in different populations.  Further sensitivity and specificity analyses are ongoing, as well as analyses focusing on the extent to which completing the Spanish-language SCQ impacts the association between ethnicity and the SCQ as many Hispanic families elected to complete the Spanish-langauge SCQ.