The study (for which I only have the abstract) applied 3 versions of the the Diagnostic & Statistical Manual (DSM) - DSM-III (1980), DSM-IIIR (1987) and DSM-IV TR (2000) - on behavioural and developmental information taken from records of just over 5,000 8-year old children suspected to have an autism spectrum disorder (ASD) from a total population of 280,423. The authors reported the numbers of cases fulfilling criteria for autism and the wider pervasive developmental disorder (PDD) according to the version of DSM used:
- DSM-III (1980) autism = 1921, PDD = 2232
- DSM-IIIR (1987) autism = 1892, PDD = 2223
- DSM-IV TR (2000) autism = 2054, PDD = 2641
They then calculated the autism and PDD prevalences (per 1000) based on the total population figure (280,423):
- DSM-III (1980) autism = 6.9, PDD = 8.0
- DSM-IIIR (1987) autism = 6.7, PDD = 7.9
- DSM-IV TR (2000) autism = 7.3, PDD = 9.4
As you may see from the figures, the prevalence of core autism seems to remain fairly constant across the various DSM versions although perhaps shows a slight upward spike using the 2000 version. The prevalence for the wider PDD category is a little more variable; indeed more so when the DSM-IV TR criteria is used. Bear in mind that this is data based on the same group of 8-year old children. I think that this quite good support for saying that at least some of the increase in prevalence is due to the changes in diagnostic criteria. Based on these figures it suggests that about 18% of the increase in the wider PDD diagnosis might be due to changing criteria (I think!).
One might ask why DSM-IV (1994) was also not included in the dataset, but I think the changes made in DSM-IV TR from DSM-IV were more cosmetic than structural (both DSM-IV TR and DSM-IV define Asperger syndrome where DSM-III did not specifically do so). Combined with the issue of diagnostic substitution covered in a previous post, you can perhaps see that at least some of the increase in prevalence is due to factors other than a 'real' increase in cases.
That is however not to say that all of the increase is due to such diagnostic factors. In the words of Prof. Sir Michael Rutter back in 2005 "a true risk due to some, as yet to be identified, environmental risk factor cannot be ruled out" (I will be coming back to Prof. Rutter in future posts). That added to an increasing prevalence and incidence of autism in certain geographical areas, leaves the door slightly ajar for maybe not just one environmental factor but quite a few, either working alone or cumulatively, spread evenly or not evenly geographically, impacting on the numbers of cases being diagnosed.
A final note. This paper appeared recently comparing DSM-IV TR and the proposed DSM-V criteria for autism. At the moment I don't have the full-text paper so cannot extensively describe and discuss. From the abstract it appears that some kind of comparison was made between those diagnosed with DSM-IV and might happen if DSM-V draft criteria were applied to the same group(?). The authors conclude in the abstract that DSM-V might be less sensitive regarding identification for those presenting with Asperger syndrome and high-functioning autism. What this might do to prevalence figures post-2013 publication is 'interesting'. I wait to see more.
Thanks IMFAR and keep up the good work.
*Reference: Rice CE, Wiggins LD, Carpenter LA, Schieve LA, Fitzgerald RT, Pedersen A, Lee LC. Variation in the prevalence of the Pervasive Developmental Disorders by diagnostic criteria. IMFAR 2011.