The utility of two autism screening instruments in an urban low-income minority population
MetadataShow full item record
IMPORTANCE: Autism spectrum disorder (ASD) represents a growing healthcare challenge that affects an estimated 1 in 68 children in the US. Early intervention can lead to improved long-term outcomes but access to these services is dependent on diagnosis. Diagnostic assessments are costly and time-consuming, making it impractical to conduct wide-scale assessments. Screening instruments such as the Social Communication Questionnaire (SCQ) and the Social Responsiveness Scale, Second Edition (SRS-2) have been developed to efficiently screen for children who should be referred for full clinical assessment. The validity of the SCQ and the SRS-2 autism screening instruments has not yet been tested in a predominantly urban low-income minority population with lower education levels and high immigrant/English as second language prevalence. Additionally, current research suggests that children with lower socioeconomic status and children in ethnic minority groups may be underdiagnosed with ASD, which underscores the importance of an effective screening method for these underrepresented groups. Minority children are typically diagnosed with ASD at a later age than white children, which leads to a delay in critical early intervention services. OBJECTIVE: We wish to examine the effects of age, gender, ethnicity, preterm birth, maternal primary language, maternal nativity, and maternal education on the validity of the SCQ-Current and the SRS-2 for pediatric patients at Boston Medical Center. A modified screening administration process was developed to account for the higher English as a second language speakers and lower-literacy rates in this population; as such, we are also examining the validity of the screening instruments when administered under this modified procedure. This study will also examine the ability of the two screening instruments to distinguish between autism spectrum disorder, intellectual disabilities, and other developmental delays. DESIGN: The SCQ-Current and the SRS-2 screening instruments were incorporated into the existing Children's Health Study, a subset of the Boston Birth Cohort. SETTING: Boston Medical Center pediatric outpatient clinics PARTICIPANTS: A total of 108 mother-child pairs completed the SCQ-Current and the SRS-2 between September 2014 and January 2015. MAIN OUTCOME MEASURES: SCQ-Current Total score; SRS-2 Total raw score; Diagnosis status of autism spectrum disorder, intellectual disabilities, or other developmental delays. RESULTS: There were six cases of diagnosed autism spectrum disorder in our sample. Both screening measures were able to distinguish between ASD and non-ASD. The SCQ had a specificity of 0.833 and sensitivity of 0.892 at an optimal cut-off score of ≥ 11 (area under the curve =0.909). Boys, preterm/low-birth weight children, children of US-born mothers, and those with other developmental diagnoses were more likely to fail the SCQ, despite not having an ASD diagnosis. The SRS-2 had a specificity of 1.0 and sensitivity of 0.961 at an optimal cut-off score of ≥ 85 (0.971). Children of US-born mothers and those with other developmental diagnoses were more likely to fail the SRS-2, despite not having an ASD-diagnosis. The measures also showed good discriminative validity (area under the curve = 0.837 for SCQ, area under the curve = 0.94 for SRS-2) for ASD vs. intellectual disability/developmental delay, though not as strong as that for the entire sample. Child's age, ethnicity, maternal primary language, and maternal education did not affect the validity of the screening measures. CONCLUSIONS: In this sample, the SCQ and the SRS-2 were valid for use in our predominantly low-income urban minority population with high preterm prevalence when combined with a modified administration process. Further research with larger sample size should be conducted to generalize these findings.