Epidemiology, Screening and Diagnosis of ASD In Adulthood

Background: Recent prevalence surveys of autism in childhood suggest higher rates than older surveys. But there is no published information on the epidemiology of autism in adults. If the prevalence of autism is increasing, rates in older adults would be expected to be lower than among younger adults. Survey methods in children can be carried out with the assistance of parents and teachers acting as informants of child behavior. This option is not available in community surveys of adults.  Adult surveys therefore require an approach that is adapted from survey methods successfully employed for studying the epidemiology of other mental and behavioral disorders such as anxiety, depression and psychosis. These methods rely on self report measures and in selected respondents at increased risk of the disorder the use of diagnostic assessments.

Objectives: To conduct an epidemiological study of autism amongst adults living in England, UK

Methods: A stratified multi-phase random sample was used in the third national survey of psychiatric morbidity in adults in England in 2007. A 20 item subset of the Autism Quotient, a 50 item self completion questionnaire, was used to screen for adults at higher risk of ASD. A second phase involved detailed assessments with the Autism Diagnostic Observation Schedule Module 4 (ADOS-4). There were 7,461 screening interviews and 618 diagnostic assessments. The ADOS-4 was also cross validated with clinical case vignettes and ADI-R and DISCO assessments conducted in a third survey phase. Respondents also provided socio- demographic and service use information. Survey data were weighted to take account of study design and non-response so that the results were representative of the household population of England.

Results: The overall prevalence of ASD in adults was estimated to be 9.8 per 1000. Prevalence was not related to the respondents’ age. Rates were higher in males, those without educational qualifications, and in those living in rented social housing. There was no evidence of increased use of services for mental health problems whereas adults with other forms of mental disorder were making greater use of treatments and services. None of the cases identified in the survey had taken part in an autism diagnostic assessment or were known to have an ASD.

Conclusions: Conducting epidemiological research on ASD in adults irrespective of whether or not in contact with clinical services is feasible. The prevalence of ASD in adults appears similar to that found in the most recent general population surveys of children. The lack of an association with age implies there has been no increase in the prevalence of ASD over time, that its causes are temporally constant and not due to a novel environmental toxic factor, and that apparent increases in prevalence in child surveys are attributable to improvements in awareness, recognition and in case finding. Use of mental health services seems less than for other mental disorders, which suggests that adults with autism are noy known to services, possibly because of under recognition and non availability of diagnostic services.