Findings from an Autism Surveillance Program in Three Regions of Canada

Background: Studies conducted over the past few decades have revealed substantial increases in the prevalence of autism spectrum disorders (“autism”), but methodological differences make comparison of the findings problematic. Ongoing surveillance is a more appropriate tool for investigating temporal changes in prevalence.  The National Epidemiologic Database for the Study of Autism in Canada (NEDSAC) was established for this purpose.

Objectives: Our objectives were to monitor the prevalence of autism in three regions of Canada—Newfoundland and Labrador, Prince Edward Island (PEI), and Southeastern Ontario—and to explore the impact of factors such as age at diagnosis and differential migration on any changes observed.

Methods: Cases of autism among children were identified through diagnostic centres, schools, and early intervention programs. Prevalence estimates were calculated for each year of the surveillance period (2003-2010 in PEI and Southeastern Ontario; 2003-2008 in Newfoundland and Labrador) and temporal trends were examined by age group (2-5, 6-9, 10-14 years) and sex by fitting log-linear models using the Joinpoint Regression Program. Age at diagnosis was compared for children diagnosed during the first and second halves of the surveillance period to evaluate the impact of this factor on changes in prevalence in the youngest age group. The estimated number of cases that moved into each region after 2003 was compared to the number that moved from the region, died, or had their diagnosis removed to determine whether, collectively, these factors likely resulted in a net increase or decrease in the numerators used to calculate prevalence.

Results: In the final year of the surveillance period, the estimated prevalence of autism per 10,000 children 6 to 9 years of age was 108.0 (95% CI: 94.2-123.3) in Newfoundland and Labrador; 99.9 (95% CI: 76.3-128.6) in PEI; and 162.5 (95% CI: 145.5-180.8) in Southeastern Ontario. The trend analysis revealed significant average annual percent increases in prevalence for the overall group of 2- to 14-year-olds ranging from 9.7% to 14.3%, but the findings varied by age group and sex: the lowest change observed was 0.9% (95% CI: -2.3-4.2) for girls in PEI, and the highest was 16.5% (95% CI: 12.7-20.4) for 10- to 14-year-olds in Southeastern Ontario.  We found no evidence that the growth in prevalence was starting to plateau except among the youngest age group in Southeastern Ontario, but that finding was likely due to methodological issues. There were no significant decreases in age at diagnosis between the first and second halves of the surveillance period, but differential in-migration may have contributed to a small portion of the observed increases in prevalence in Southeastern Ontario and PEI.

Conclusions: The prevalence of autism increased significantly among all age groups, yet the factors we examined accounted for only a small portion of the observed increases. Accordingly, we cannot rule out the possibility of a true rise in incidence, particularly given the lack of evidence of a leveling-off of prevalence in the younger age groups. Continued surveillance is needed to explain observed increases in prevalence.