Most children with autism spectrum disorder (ASD) are not diagnosed until the age of 4, thus missing the opportunity for early intervention. The objective of this study was to investigate the feasibility of an early screening program for ASD applied during well-child visits in a community-based sample. The study lasted for 4 years and was divided into two stages. Stage I involved the implementation of the basic screening model in 2014. Toddlers received level 1 screening via section A of the Chinese-validated version of the Checklist for Autism in Toddlers (CHAT-23) during 18- and 24-month well-child visits in Xuhui District, Shanghai, China. Screen-positive children were referred to receive section B of the CHAT-23 for level 2 screening, and those still screen-positive were referred to undergo diagnosis and evaluation. Stage II involved the implementation of the improved screening model from 2015 to 2017 with the following modifications: (a) an added observational component in level 1 screening; (b) telephone follow-ups with the screen-positive families; and (c) dissemination of information on ASD to families. The results showed that 42 of 22,247 screened children were diagnosed with ASD. The ASD diagnosis rates were 0.1% in Stage I and 0.21% in Stage II. The screen-positive rate and the show rate of referral for level 1 screening increased by 76.92% and 58.43%, respectively, in Stage II compared to Stage I. Our results suggest that with appropriate logistic support, this two-level screening model is feasible and effective for the early screening of ASD during well-child visits. Autism Res 2018, 11: 1206-1217. © 2018 International Society for Autism Research, Wiley Periodicals, Inc. LAY SUMMARY: Difficulty in the timely identification of autism spectrum disorder (ASD) results in missed opportunities for many ASD children to receive early intervention. In this study, we established an early screening model for ASD among children aged 18-24 months in the community by relying on the three-level child healthcare system in China. The results showed that this model can effectively identify and diagnose ASD in children at an early age and thus enable early intervention.
