Having spent the past 6 years on what have become the Depressive Disorders and Bipolar and Related Disorders chapters in DSM-5, I have been invited to reply to Dr. Parker's comments.

I will not try to respond to every issue Dr. Parker has raised, but I will focus on the context and rules of procedure that determined the content of DSM-5, which might help in understanding how DSM-5 became what it is.

Three contextual factors directed the production of DSM-5 that might cast some light on what some would see as its ‘limitations’.

One prominent factor was the hope that enough research existed to allow us to address the heterogeneity of mood diagnoses such as major depression. This issue was addressed by Dr. Parker in his critique that major depression has been reified instead of being considered a ‘domain’ diagnosis, based on unchanging criteria, supported by a great many studies—and questioned by a relative few. In the beginning of the DSM-5 process, it was thought that there might be adequate support for dimensional diagnosis or biological findings to support the validity of some diagnoses. Examination of existing genetic, imaging, and clinical data led to the realization that we on the DSM-5 mood workgroup had been born 20–30 years too early and that the current data available were not adequate to make these advances.

The second factor shaping DSM-5 process was the American Psychiatric Association, Board of Trustee's decision that any major change would require the approval of a Scientific Review Committee (SRC). This meant that though not everything in DSM-IV was strictly evidenced-based, to make any substantial changes, evidence from the literature, as well as the reason for changes and a discussion of possible unintended negative consequences, had to be submitted to the SRC and be given a ‘passing’ score to be included in DSM-5. This lead to a conservative approach to changes and in the case of a lack of coverage in the literature, the inability to make ‘wished for’ changes.

The third contextual condition is related to the first, discussed above. We were acutely aware that this DSM was being put together on the historical verge of a paradigm shift in our thinking about diagnosis. As Loscalzo and Barabasi state in their article ‘Systems Biology and the Future of Medicine’ [2], ‘The 19th century – Oslerian formalism for human disease links clinical presentation with pathological findings. – While there has been continual refinement of the pathological markers used for correlation (e.g., biochemical measurements, immunohistochemistry, flow cytometry, and more recently, molecular pathogenic analyses of expressed genes), This paradigm has been helpful to physicians as it establishes syndromic patterns that limit the number of potential pathophenotypes they may need to consider. While quite useful at an earlier eras, classifying disease in this way vastly overgeneralizes (reifies?) pathophenotypes —and — cannot be used to individualize disease diagnosis or therapy.’ The authors go on to point out ‘it is hardly surprising that these conventional pathophenotypes are far too limited to be useful in the postgenomic era.’ These authors use the example of the Mendalian disease, sickle cell disease that is caused by a single point mutation at a known position of a known gene. Despite the specificity, this simple phenotype does not yield a single diagnostic picture, but presents in six different pathophenotypes dependent on environmental conditions and interaction with other as yet unknown genes.

In approaching the DSM process with its reliance on clinical symptom—derived phenotypes, we are in the position of ‘conscious ignorance’ formulated by neurobiologist Sidney Firestone in his recent TED lecture. We know enough to know that the traditional DSM approach is obsolete, but we do not know enough to articulate the paradigm shift we know will be necessary. We see this in treating refractory mood disorders—we fire a sequence of ‘arrows’, medications with different mechanisms of action from our ‘quiver’ until with persistence, and most of our patients ultimately respond. Right now, with major depression as our target, we are shooting in the dark, hoping to alter a (currently unknown) mechanism that will reverse the symptoms of the suffering patient. A recent example is the dramatic, but time limited effect of intravenous ketamine in about half of patients that receive it.

There is not currently enough knowledge to get beyond that—one of the reasons that DSM-5 is not titled DSM-V. In the computer–internet age, we can update the DSM-5 to 5.1, 5.2 etc. as knowledge develops without having to wait another 12–18 years for a revision.

If we develop the knowledge for ‘personalized medicine’, much as has been recently done in cancer chemotherapy, where location and tissue source is less important than the specific genetic drivers of malignancy, we would not have to debate whether specifiers are positioned as diagnostic subtypes or whether specifiers are clinically justified [3].

So, OK, maybe the whole concept of DSM definitions of psychiatric disorders is obsolete—maybe it has served its purpose, and now it is time to move on. This is my personal opinion, but in the meantime, we need something—as much as I would have liked it to be more—knowing we are just barely on this side of the edge of history—until we can make the next advance. Until that knowledge is developed, we can debate how to make something that is very incomplete, serves us as best as possible.

As to Dr. Parker's detailed critique, it is appreciated. In many, but necessarily not all cases, contextual limitations would have precluded the clarification that he has suggested.


  1. Top of page
  2. References