Cumulative Incidence and Prevalence of Autism Spectrum Disorder

Abstract

Importance Between the 1980s and 2020s, reported autism spectrum disorder (ASD) prevalence has increased markedly, raising public concern. However, most estimates rely on administrative or service-based data vulnerable to diagnostic shifts, service-related ascertainment, and policy-driven incentives. Objective To assess whether 7-year cumulative incidence and prevalence of ASD at age 7 years changed across consecutive birth cohorts. Design, Setting, and Participants Prospective population-based screening and diagnostic assessment was conducted in 12 consecutive birth cohorts in South Korea among children born from January 1998 to December 2010 and ascertained at age 7 years between January 2005 and December 2017. All children aged 7 years entering elementary schools constituted the target population. Among more than 62 000 children, 2077 screened positive and were invited for detailed evaluation, and 881 completed comprehensive diagnostic assessments. Data were analyzed from January 2024 to July 2026. Exposures Birth year and service-use status defined as any history of prior diagnosis or intervention. Main Outcomes and Measures Populationwide standardized screening used the Autism Spectrum Screening Questionnaire; diagnostic assessments, including the Autism Diagnostic Observation Schedule, Autism Diagnostic Interview–Revised, cognitive testing, and record review, were administered to screen-positive children. Clinical consensus best-estimate diagnoses were determined according to Diagnostic and Statistical Manual of Mental Disorders , 5th Edition , criteria. Results Across consecutive birth cohorts, both 7-year cumulative incidence and prevalence estimates of ASD at age 7 years remained stable, ranging from 1.9% (95% CI, 1.5-2.4) to 3.2% (95% CI, 2.7-3.8) for incidence and from 2.0% (95% CI, 1.5-2.6) to 3.4% (95% CI, 2.7-4.0) for prevalence, with no statistically significant increase or decrease. Female children with ASD were disproportionately represented among previously undiagnosed cases identified in general education settings and exhibited distinct phenotypic profiles and substantial unmet clinical needs. Conclusions and Relevance The findings from this cohort study suggest that detection-related factors may contribute substantially to reported increases in ASD occurrence estimates. Interpretation of changes in ASD occurrence estimates should account for methodological differences, changing service-use patterns, and ascertainment context in service-based surveillance systems.