Autism Under the Proposed New DSM-5 Criteria

As the DSM5 moves closer to its anticipated release next year, several areas are generating widespread debate and discussion both in the public and in more academic circles.  One diagnosis certainly getting its share of attention is Autism.  Under the proposed new guidelines, a single Autism Spectrum Disorder (ASD) will replace the different sub-diagnoses of Asperger’s Disorder and PDD-NOS, and the former three part criteria will be reduced to two: social communication deficits and fixated interests and repetitive behaviors (social interaction and communication were separate in DSM-IV). 

Just the language changes themselves have triggered some concern, particularly among those who prefer terms like Asperger’s Disorder that convey some distance between it and more classic Autism.  Generating even more controversy is the worry that the new criteria might trigger large changes in who will and will not meet criteria for an ASD in the future.  While some are distressed that more relaxed criteria will label higher and higher percentages of children with a major psychiatric disorder, others fear that more stringent criteria might cause some children to lose important services. 

The latter concern was heightened by a new study that examined the specificity and sensitivity of the new DSM5 criteria as applied to a group of children with autism as defined previously.  The authors reported that while the specificity of the new criteria was excellent (i.e. the new criteria resulted in very few false positives), the sensitivity (i.e. the false negative rate) was much reduced, resulting in fewer ASD diagnoses particularly among those with higher cognitive function.  Overall, the authors reported that 60.6% of children with a previously defined ASD would meet criteria under the new guidelines.  

Adding to the drama are some additional characteristics of the authors and the study, including the fact that the senior author who was previously on the committee tasked to make the new criteria later resigned and the fact that there was a premature release of the study to the public media prior to publication.  Indeed, the study was much more in the public spotlight months ago after articles in the New York Times and elsewhere than it is now with the paper’s official publication this month. 

In the same journal issue, members of the DSM-5 Workgroup on Neurodevelopmental Disorders sharply pointed out some important flaws in the study, such us the study’s use of 20 year old data onto which the authors had to apply DSM-5 criteria that were never explicitly asked as such.  The study was also not representative enough to make the kinds of sweeping conclusions that have been reached.  The Workgroup notes that many of their criteria are less stringent than the DSM-IV items and cites more recent studies indicating a superiority of the new model. 

One area where there is agreement is in the need to work further on the issue to get it right or as good as it can get given the limitations of the science.   Until then, more alarmist concerns that either everyone or no-one will meet criteria for an ASD starting next year seem unwarranted. 


McPartland JC, Reichow B, Volkmar FR.  (2012). Sensitivity and specificity of proposed DSM-5 diagnostic criteria for autism spectrum disorder. J Am Acad Child Adolsc Psychiatry 51(4):368-383.

Tags: , , ,

Leave a Reply

You must be logged in to post a comment.