Thursday 17 November 2011

Starring autism in DSM-5

It has been several years in the making but like a Hollywood blockbuster come May 2013, DSM-5 or DSM-V if you prefer, will be heading to a psychologist, psychiatrist or combined CAMHS team near you, rated... well I will leave that decision to you.

There have been quite a few debates about DSM-5. More generally both the American Psychological Association and the British Psychological Society (BPS) have voiced some concern over several issues with the proposals which have been summarised in an open letter. The main factors highlighted include: the rapid expansion of the diagnostic system, a perceived lack of scientific rigor and review and an overly-complex dimensional aspect to many conditions within the DSM-5. The BPS is also particularly concerned about the apparent lack of recognition of social factors such as poverty as being pertinent to a person' psychological health and wellbeing and the impact of stigma. I covered something similar to this in a recent post on the current woes of Greece.

We will see what becomes of these concerns voiced by such learned societies.

With autism spectrum conditions specifically in mind, I have previously covered a few topics of potential interest to DSM-5 on this blog. There have for example, been subtle clues to how autism will pan out in the new DSM revision for quite a few years now as evidenced by things like changes to the algorithm used as part of ADOS and papers looking at historical recall with regards to the cut-off ages for presentation of the autism.

Another clue to the changes to autism diagnosis came quite recently following the publication of a paper by Catherine Lord and colleagues* on the lack of consensus about best-estimate diagnostic sub-groups labels applied across different trained groups.

The paper by Lord (who has fairly recently moved jobs) et al has received quite a lot of press, along the lines of 'Autism or Asperger's? It might depend on your doctor' reiterating what has perhaps been known about for many years: physicians are artists and brush stokes vary from physician to physician. The latest paper looked at just over 2100 children diagnosed with autism via ADI-R and/or ADOS spread across 12 participating sites. Based on various pieces of information covering various autism- and cognitive-related measures, best-estimate diagnostic labels were compared across the sites and lo and behold, the artists, sorry physicians, were not in total agreement about labels. So a child's final diagnostic label is, to a degree, dependent on who synthesised the pertinent material and gave the diagnosis.

I note that quite a few 'I told you so' soundbites have followed this study. I pass no comment on these except for the fact that DSM-IV, and its modifications, has been with us for 17 years so far and perhaps therefore the question of why it has taken so long to do a study like this should be asked. Did no-one ever think that this should have been part of the field testing for DSM-IV?

I digress. The study has been used as evidence for the assignment of a dimensional description to the triad (sorry dyad) of symptoms as per the DSM-5 recommendations. There is some growing support for the options laid out in DSM-5 from other sources; so imaging studies, more imaging studies and co-morbidity research. Despite this, the research is not all one-way; not at all. I don't think this will be the last piece of research to be published pertinent to the important diagnostic changes put forward in DSM-5.

Finally, I do think we should try and keep in mind two important things: (i) autism or whatever variant detailed is not just a label; it is not just a collection of symptoms gathered under a few sentences of descriptions, it is real life and as complicated as 'not autism', and (ii) changes to those labels need to be done for the right reasons; so if it means that people will better receive the help, support, services, etc they need/want under DSM-5, targeted to that persons strengths and weaknesses, so be it. If it however means that it is just easier to categorise people for a research study or investigation based on where they appear on the autism spectrum, I would perhaps suggest that we should be investing more in suitable research tools for this purpose.

Speaking of physicians as artists, have a read through this recent article** (click the 'Article as PDF' option in the right sidebar to download it, not forgetting to scan for viruses first).

* Lord C. et al. A multisite study of the clinical diagnosis of different autism spectrum disorders. Archives of General Psychiatry. November 2011.
** Gupta VB. & Lauffer D. Should the diagnosis of autism be made only on the basis of a standardized test? Journal of Developmental & Behavioral Pediatrics. November 2011


  1. Some great points raised.

    The new Lord et al. paper is certainly interesting - and quite sobering I think. However, the way it's been presented in the media gives the impression that the same kids were assessed by different physicians, when in fact what they showed was that different kids with similar scores on standardized tests (particularly the ADOS) ended up with different diagnoses depending on which centre was giving the diagnosis. But the clinical diagnoses are made on much more than the number of boxes ticked on the ADOS algorithm. So to conclude that physicians get it wrong (or are at least in strong disagreement with one another), the authors have to assume that the ADOS is a perfect instrument - which it very clearly isn't.

  2. Thanks Jon. The Gupta paper listed at the foot of the post is in agreement with you on this one in that ADOS (and ADI) are put forward as 'gold-standards' but the standard is relative to the fact that autism as a label/description is still very much a diagnostic maze (not even getting me started on the confounders like comorbidity, etc). MJ @ Autism Jabberwocky did a nice post not so long ago which covered ADI and some of the issues there:


Note: only a member of this blog may post a comment.