Cumulative Incidence of Autism Spectrum Disorders in a Total Population Sample of 5-Year-Old Children

Background: Recent reports from Center for Disease and Control and other research suggest an increase in Autism Spectrum Disorder (ASD). Despite a growing number of publications reporting on the ASD prevalence, however, it remains unknown if there is a true increase in ASD prevalence. An increase in the prevalence and the incidence of ASD can be attributable to: 1) better ascertainment; 2) broadened ASD diagnostic criteria; and, 3) a true rise in ASD incidence. One way to overcome this challenge is to examine the ASD incidence, another measure of ASD frequency and compare the incidence annually. However, only few studies have examined the ASD incidence mainly due to the methodological challenges in estimating incidence; more specifically, the difficulty in identifying the time of disease onset in ASD research.

Objectives: To determine whether there is a true increase in ASD prevalence by investigating the cumulative incidence of ASD annually.

Methods: Using a total population sample in Hirosaki city in Japan, all 5-year-old children in the catchment area underwent the screening phase annually from the year 2013 to the year 2016. Children who screened positive were invited to comprehensive assessment at the University Clinic, including child and parent interview, behavioral observation, cognitive testing, and motor function testing. Each case was discussed in a multidisciplinary research team, and the best estimate diagnosis was determined based on findings from screening tools and diagnostic assessment, as well as clinical judgement. The same screening tools and diagnostic testing as well as the same diagnostic criteria (DSM-5) were used in all study years to minimize the impact of changing criteria on ASD ascertainment and of changing diagnostic criteria on prevalence and incidence estimates. We estimated cumulative incidence of ASD up to 5 years of age for each study year by dividing the number of children diagnosed with ASD who were born in the catchment area by the number of the birth cohort in the catchment area each study year.

Results: Caregivers of 3,954 children completed and returned the screening packet for a participation rate of 78.8%. Among 773 children who were screen-positive, 559 children underwent the assessment, and 87 children were eventually diagnosed with ASD. The cumulative incidence of ASD up to 5 years of age for the total study years was 1.31% with the 95% CI of 1.00 – 1.62. Although the 5-year cumulative incidence increased from the year 2013 to the year 2014 and from the year 2014 to the year 2015, further analysis, using a generalized linear model revealed that there were no significant linear trends in 5-year cumulative incidence over the study years 2013 – 2016.

Conclusions: Taking advantage of the sequential study design allowed us to compare the annual 5-year cumulative ASD incidence and successfully demonstrate that there was not a true rise in ASD incident cases over the 4-year study period in the study catchment area.