Parent-Teacher Agreement on An Autism Screener in An Underserved Preschool Population

Background:  The importance of early diagnosis of autism spectrum disorders is well-established. Despite published guidelines for early screening for ASD, screening is not widespread in healthcare settings. This is especially true for impoverished children who may not have a regular source of care. Preschool and daycare programs, therefore, may be an important additional venue for screening. Review of the literature suggests that utility of preschool teacher screenings has not been explored. This study compared, in a traditionally underserved sample, parent and teacher responses on commonly used autism screening tools, with subsequent comparison to clinical diagnosis of autism. 

Objectives: The objectives are to 1) measure agreement between parents and teachers on commonly used autism screening tools, 2) explore concordance between screening results and subsequent clinical diagnosis, and 3) identify factors that are associated with concordance. 

Methods: Six cities in New Jersey with large low-income, minority populations were selected. Preschools in target cities were invited to participate in the study.  Preschool staff received training on the symptoms of autism. Parents and teachers were asked to complete the Modified Checklist for Autism in Toddlers (M-CHAT) and/or Social Communication Questionnaire (SCQ). Children screening positive on either instrument received follow-up phone interviews to clarify responses. Those who screened positive on the follow up interview were offered a clinical evaluation that included Autism Diagnostic Observation Schedule (ADOS) and cognitive screening. Results were analyzed to determine parent and teacher agreement between raters for each instrument and pattern of agreement between screening results and clinical diagnosis.

Results: Results for both parent and teacher M-CHATs were obtained for 190 children. Completed results were obtained for an additional 405 children having both parent and teacher SCQs. Screened negative rate by parent report was 81% for the M-CHAT and 92% for the SCQ. Screened negative rate by teacher report was 79 % for the M-CHAT and 93% for the SCQ. Agreement between parent and teachers was 76% for M-CHAT; 90% for SCQ. For parents and teachers, positive predictive validity (PPV) of M-CHAT was 44% and 58%, respectively; PPV of SCQ was 40% and 41%; sensitivity of M-CHAT was 50% and 100%; sensitivity of SCQ was 36% and 78%; specificity of M-CHAT was 50% and 37.5%; and specificity of SCQ was 65% and 37.5%.

Conclusions: Relative merit of these screening tools and influence of demographic factors will be discussed. Findings suggest that parents in this underserved population poorly identify cases at risk for autism, whereas preschool teachers are good at identifying cases at risk for ASD but not as accurate identifying those children not at risk. Thus, including both parent and teacher screening reports increases the accuracy of finding cases of ASD in a preschool population.