Abstract

Aims: To determine the prevalence of autistic spectrum disorder in a geographically defined population for children receiving services and compare this to the estimated prevalence based on a two source “capture–recapture” technique previously employed in biological populations to give a “true” prevalence with full ascertainment.

Methods: Information on gender, age, and postcode sector was determined from nine different datasets maintained for children with autistic spectrum disorder and point prevalence was calculated. Data from the diagnostic services and the Special Needs System were entered into the two source “capture–recapture” calculation.

Results: Of a total population of 134 661 under 15 year olds resident in Lothian in southeast Scotland, 443 were known to autism services, with a point prevalence of 32.9 per 10 000 (95% CI 29.8 to 36.0). The estimated prevalence using a capture–recapture method was 44.2 (95% CI 39.5 to 48.9), which suggests that 74% of affected children were registered with services in some way. The age distribution was similar to that of the background population under the age of 12 years and there was no indication of a rising prevalence. The ratio of boys to girls was 7:1.

Conclusions: The prevalence of autistic spectrum disorder in a geographically based population employing two source capture–recapture analysis is comparable to that quoted for the best active ascertainment studies. This technique offers a tool for establishing the prevalence of this condition in health service populations to assist in planning clinical services.