37

This article provides a review of substantial findings and methodological issues in epidemiological studies on autism. Studies published since 2000 were reviewed and a large degree of heterogeneity of methods across surveys can be observed. Prevalence estimates vary widely, with a range of prevalence from 0.7 % to 1.5 % being consistent with recent, well-designed studies. Factors that explain time trends in prevalence are examined, including changes in diagnostic concepts and criteria and improved awareness and case ascertainment. Finally, we review how factors such as social class and ethnic minority status affect prevalence in subgroups.