Identification and Diagnosis of Autism Spectrum Disorders: An Update

Greg Pasco


Pediatr Health. 2010;4(1):107-114. 

In This Article


Lotter's original prevalence study (published in 1966)[35] estimated the prevalence of autism to be approximately four per 10,000 children. Four decades later, the Special Needs and Autism Project (SNAP) suggested that the prevalence of ASDs is 116 per 10,000 children.[36] This latter finding is consistent with numerous studies that have been published in the last couple of decades that have estimated the prevalence of ASD to be in the region of 1%.[37] See Box 3 for a description of the changes in measured prevalence of ASDs over time. Whilst there have been many changes in the way that autism has been defined and conceptualized over the past four decades, particularly in relation to Asperger syndrome and the other variants of autism, there has been consistent controversy about whether the incidence of autism is actually on the increase, and what the causes of the autism 'epidemic' may be.[38] In particular, the belief that the measles–mumps–rubella (MMR) vaccination was a potential cause of autism created enormous concern in many countries. These claims, largely based on a now discredited study, have been contradicted by numerous subsequent studies investigating the relationship between changes in prevalence and the introduction or withdrawal of MMR vaccination programs.[39]

Diagnostic estimates in the SNAP study were subdivided into categories ranging from the strictest and most restrictive definitions for autism to the broadest definitions for the whole autism spectrum. The prevalence for ICD-10 childhood autism was just under 39 per 10,000 children. A subsample of this group was subsequently rated according to the criteria used by Lotter, with approximately 80% found to meet his original definition of autism,[40] suggesting that there may well have been an increase in prevalence.

Investigations have focused on a number of potential explanations for the measured increases in the prevalence of autism. King and Bearman argue that with regard to autistic disorder, diagnostic substitution (individuals previously diagnosed with mental retardation and subsequently given a sole diagnosis of autism) and accretion (where a diagnosis of autism is added to the previous diagnosis as a comorbid condition) accounted for approximately a quarter of the increase in prevalence in California, USA.[41] The authors suggest that diagnostic change was most likely to occur in a year during which diagnostic practice changed, whether due to the publication of new diagnostic criteria or to changes in local policy or service provision. A study of the diagnosis of all ASDs in Western Australia[42] reached similar conclusions in relation to observed increases in prevalence. The authors speculate that children with borderline mental retardation may now be more likely to be diagnosed with autism than previously.

In a systematic review of prevalence studies, Williams et al. conclude that a number of factors account for the variation in prevalence estimates.[43] These factors include: the diagnostic framework used; the year in which cohorts were investigated; the method of ascertainment (prospective studies yield higher figures than retrospective studies); the region where the study was conducted (e.g., Japanese studies have reported higher figures than those conducted in the US, and urban areas were associated with higher prevalence than rural areas[43]); and the age of participants. Chakrabarti and Fombonne have reported the findings of prevalence studies involving two consecutive birth cohorts ascertained by identical methods from the same geographical area of England.[44,45] They conclude that there was no evidence of an increase in incidence of pervasive developmental disorders, although they acknowledge that in comparison to earlier studies, measured prevalence may have increased as much as threefold.

Rutter concludes that it is not possible to know whether there is a true increase in the incidence of autism, but that much of the increase in the measured prevalence of all ASDs is associated with the increased awareness of Asperger syndrome and autism in people with IQ measurements in the average range.[46] Whilst dismissing the MMR vaccine as a causative factor, Rutter suggests that risks associated with other prenatal or postnatal toxins as well as demographic factors such as a rising age of parenthood should not be dismissed. The complexity of conducting a definitive epidemiological study, and the difficulties of comparing current measures of prevalence with earlier data, have largely confounded attempts to answer the simple question of whether the true incidence of autism is on the increase. The consensus is that the level of increase remains unexplained and that there is a need for more and better studies.[38,40,41,43,47] The ideal model for such studies would be longitudinal and involve consistent case definition and ascertainment procedures within the same geographical area. The Autism and Developmental Disabilities Monitoring Network in the USA employs an ongoing, multiple-source, case-ascertainment procedure across a number of states and reports that this approach has provided prevalence estimates with adequate levels of representativeness, sensitivity and reliability.[48]


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.