the need for a thoughtful and comprehensive diagnostic assessment (discussed below)—this is particularly true for preschool children, for whom issues of diagnosis can be complicated.


Epidemiological studies of autism have important implications for both research and clinical service, for example, through helping to plan for the need for special services and selecting samples for research studies. Fombonne (1999) has recently summarized the available research on this topic and systematically reviewed more than 20 studies conducted in ten countries. As he notes, although important details were sometimes lacking in the studies, the total population base included in the review was approximately 4 million children surveyed.

Studies have typically employed a two-stage design, with an initial screening followed by more systematic assessment. Complexities in interpreting the available data include variations in approaches to diagnosis of autism and differences in screening methods. In the studies surveyed, approximately 80 percent of individuals with autism also exhibited mental handicap (i.e., mental retardation). Studies have also consistently identified more boys with autism than girls (three to four boys for every girl). In addition, girls with autism are more likely than boys to also exhibit mental handicaps.

The epidemiology of autism has recently become quite controversial. In the United States, increased demand for autism-specific services (Fombonne, 1999) has drawn attention to growing numbers of children with the educational categorization of autism. Large, systematic epidemiologic studies have reported increases in prevalence from the estimated rate of 2–5 per 10,000 in the 1970s to 6–9 per 10,000. Fombonne (1999) considers a rate of 7.5 per 10,000 to best reflect the result of studies conducted since 1987. Those studies also report a rate of 12.5 per 10,000 individuals for atypical autism/pervasive developmental disorders, producing an overall rate of about 20 cases per 10,000. Rates for Asperger’s disorder, excluding individuals who also met criteria for autism, were low, at 1–2 per 10,000. Because the studies did not consider individuals with less-pronounced variants of autistic spectrum disorder, it is possible that the figures for atypical autism/pervasive developmental disorder and Asperger’s disorder are underestimates.

Two simple reasons explain the difference in current and historical rates: more complete diagnoses and a broader definition of autistic spectrum disorders (Fombonne, 1999). However, there are a number of recent studies, most with small samples, and several reports from school systems that found even higher rates of autism (Centers for Disease Control and Prevention, 2000; Arvidsson et al., 1997; Baird et al., 2000; Kadesjoe et

The National Academies of Sciences, Engineering, and Medicine
500 Fifth St. N.W. | Washington, D.C. 20001

Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement