Quantitative Evaluation of Sociologic Factors That Can Lead to Apparent Increases in Autism Prevalence

Background: The prevalence of autism (‘autism spectrum disorder’) has risen dramatically from less than 1 per 2500 before 1980 to 1 per 91 children in 2007.   The autism literature commonly cites three sets of sociologic factors, namely changes in diagnoses, enhanced professional and parental awareness, and Special Education funding incentives, as being largely responsible for the increased autism prevalence.   Previous work quantifying diagnostic changes suggest that this cause does not explain the rise in autism prevalence in the US. The quantitative validity of the other sociologic factors has not been evaluated. 

Objectives: To quantitatively measure the contributions of increased awareness and Special Education funding to the rise in ASD prevalence.

Methods: A review of autism prevalence literature, publicly available documents such as the US Statistical Abstracts, documents from Department of Education,  Yahoo group sites, and other relevant databases was conducted.    US autism/ASD prevalence rates from peer-reviewed publications were combined and averaged if more than one prevalence was published for a given birth year.   When prevalence was measured for a range of years, point prevalence was obtained by taking the midpoint of the years.  Professional awareness was measured through historical counts of the number of professionals who can diagnose autism and the number of peer-reviewed publications related to autism, as listed in Pubmed.   Peer reviewed autism publications were normalized to the total number of annual peer reviewed publications, in order to account for confounding variables such as increased ease of publishing due to word processing advances or expanded numbers of scientific journals.    Non-professional awareness (e.g., parental) was measured through Yahoo autism groups usage. Software was written to automatically download Yahoo group websites and extract the number of messages in Yahoo groups devoted to autism.    The number of messages was normalized to the total number of Yahoo health (but non-autism) group messages in order to account for general increased computer usage and Yahoo groups.    The history of autism funding by the Department of Education was reviewed and the funding amounts normalized to spending on general education.

Results: The US autism prevalence had a fold change of approximately 37 from birth years 1977 to 2000. The normalized number of autism diagnosing professionals had a fold change of 2.48 from 1977 to 2002.  The normalized number of autism publications had a fold change of 3.3 from 1977 to 2002. Increased parental awareness was not significant until the late 1990s. Federal funding for Special Education was disbursed starting only in 1995, and had a normalized fold change of 2.24 from 1995 to 2006.

Conclusions: Increases in professional and parental awareness and increase in federal funding for autism do not adequately explain the rise in autism prevalence.