Dan Stein offers us a ringside view of multiple considerations that went into recent revisions of just one subsection of psychiatric nosologies (in DSM, ICD-11, and RDoC): the subsection covering obsessive-compulsive and related disorders (OCRDs). These considerations include results from imaging neuroscience, results from animal models, genetic data, results from field studies, clinical data and experience, and patient advocacy. In addition, he finds historical contingency, or arbitrariness, in some of the outcomes. I find the story fascinating and philosophically rich. In my comments, I draw attention to five of the themes and develop them a little further. I conclude with some observations about conservatism in revision of psychiatric nosologies.
9.2 Prematurity of the science
Many psychiatric researchers hoped that the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) would incorporate recent results of genetic and neuroimaging studies, going beyond the “cluster of symptoms” approach that characterized DSM-III and DSM-IV to discern more of the etiology of psychiatric disorders. In the area of OCRDs, there have been intriguing results from neuroimaging. In particular, it looks as if the OCRDs are distinguished from anxiety disorders in that the activation is in the forebrain rather than in the amygdala. These studies, which suggest the anatomic outlines of an etiology (although not the etiology itself) were the basis for classifying the OCRDs separately from the anxiety disorders. However, other empirical results do not support this etiology. For example, cognitive-behavioral therapy and selective serotonin re-uptake inhibitors (SSRIs) are effective treatment modalities for both sets of disorders, and the disorders have considerable comorbidity. These results suggest a common, or at least overlapping, etiology, and could reasonably have been used to resist separating the OCRDs from the anxiety disorders. (The possibility of more than one reasonable conclusion is something I will address in section 9.6.) It is not possible to predict how the different streams of evidence—from neuroimaging, genetics, successful treatment, and comorbidity—will develop over time. At this time, new empirical results justify only provisional and modest changes in nosology.
9.3 Pluralism of nosologies
Although most of the focus is on the DSM, Stein describes the situation for three different nosologies: the DSM-5, the eleventh edition of the International Classification of Diseases (ICD-11), and the Research Domain Criteria (RDoC). DSM-5 is the classification produced by the American Psychiatric Association and is intended for both researchers and clinicians; ICD-11 is internationally and clinically oriented; and RDoC is a proposed dimensional framework for research that is designed to lead to identifying biomarkers and causal etiologies for mental disorders. (RDoC was developed with the support of Thomas Insel at the National Institute of Mental Health, and has an uncertain future now that Insel is no longer leading that organization.) No one wants the classification systems to diverge too much, because this would lead to confusion over applying the results of research to clinical practice, and vice versa. But small differences are manageable, and serve as a reminder that our classification systems do not (yet) carve nature at the joints. As an example, olfactory reference disorder is having a better reception in ICD-11 discussions than it had in DSM-5 discussions.
9.4 The role of historical contingency
Classification outcomes are dependent on the timing of the committee discussions, the salience of anecdotal experiences, and the influence of authoritative experts. None of these closely track new evidence, and so the result can be a somewhat arbitrary outcome due to what some philosophers of science call “external factors” (I have called them “non-empirical decision vectors” in Solomon 2001). An example is hoarding disorder, which according to Stein benefited from being discussed early in the DSM process, as well as from the support of an influential member of the Scientific Review Committee.
9.5 Patient advocacy
Tourette’s syndrome has traditionally been classified among the OCRDs. Some patients and their advocates have preferred that it be classified as a neurological, rather than a psychiatric, disorder. This is because they do not wish to be affected by the stigma sometimes associated with an OCRD psychiatric diagnosis. Dan Stein points out that in resisting the OCRD diagnosis, these patients and their advocates are implicitly reinforcing stigmatization for those who have OCRD psychiatric diagnoses. (This is what Hilde Lindemann (in Nelson 2001) calls a “hostage taking” identity, and in Chapter 23 in this volume I explore a similar situation with the Asperger’s diagnosis.) DSM-5 has gone with the wishes of these patients and their advocates, classifying Tourette’s with neurodevelopmental disorders such as attention-deficit hyperactivity disorder (ADHD) and autism. On the other hand, ICD-11 classifies Tourette’s in both categories—OCRD and neurodevelopmental disorders. Tourette’s has considerable comorbidity with both obsessive-compulsive disorder and ADHD, and has been effectively treated with medications and cognitive-behavioral therapy. Thus there is evidence in favor of keeping the traditional classification.
9.6 “An integrative approach”
Dan Stein describes the overall situation when revising psychiatric nosologies as one in which both evidence and values play a role and need to be integrated in an overall decision about any changes. He also uses an overlapping dichotomous classification—which he calls “classical” and “critical”—to describe the joint influences of science (corresponds with “evidence”) and pragmatic choices (corresponds with “values”). Another dichotomous classification covering similar (but not the same) same ground is “validity” versus “utility.” As a first approximation, I think that framing the situation in terms of the integration of factors that fall into these binaries is helpful. But it is only a first approximation. The binary analyses obscure the fact that often it is empirical/scientific factors that conflict with one another (e.g., results of neuroimaging and results of clinical trials with behavior therapy), as well as pragmatic factors/values that conflict with one another (e.g. some patients’ values may conflict with those of some psychiatrists, or some clinical demands may conflict with reimbursement practices). In addition, the three sets of binary categories used—classical/critical, validity/utility, and evidence/values—do not precisely coincide with each other. For example, evidence may tell against a “classical” classificatory scheme, and “values” may conflict with “utility.”
Moreover, use of the language of “integration” of different factors relevant to classification does not tell us how the integration is done and in particular whether it is possible to do the integration in more than one way. It is likely that there is often more than one apparently reasonable decision, and unrelated factors (such as belief perseverance, or wish to make an impact, or the influence of powerful individuals) can play a deciding role. This is a problem that is not unique to psychiatry; it comes up in any area of science policy in which the evidence is complex,multimodal, and sometimes contrary (Stegenga 2011). Science and technology studies researchers Harry Collins and Trevor Pinch (1993) have argued that, when we study science and technology, “the mess is the message.” Stein is right to note that “nosological revision is a messy business.”
Finally, some comments about conservatism. DSM-5 ended up being much closer to DSM-IV than was expected or hoped for. This is in part because evidence for a new nosology has not yet reached critical mass. But it is also because of systematic processes that work against change. Every change to the DSM has costs in terms of education, communication, practice change, administrative overhauls, etc. Projected benefits from making a change need to outweigh the costs. Another factor is that the experts invited to serve on consensus committees typically participate in decisions about more than one classificatory system: for example, there is some overlap between those who serve on DSM committees and those who serve on ICD committees. Stein even mentions the role of a joint team of those leading DSM-5 and ICD-11 in settling on a common “metastructure.” The deliberative processes of those committees are thus not independent, and the overlap of experts further reduces any potential for difference. If the processes were independent—and especially if they were concurrent—more difference would be likely.
Kenneth Kendler (2013) has pointed out that that there is one factor working against conservatism: the common desire among people who participate in these work groups to show that they have made a difference. Leaving a nosology the way it was before can make it look as if no work was done (even though much work may have been done). Is this factor an adequate corrective to the systematic processes resisting change?
Collins, H. and Pinch, T. (1993). The Golem: What everyone should know about science. Cambridge, UK: Cambridge University Press.Find this resource:
Kendler, K. (2013). A history of the DSM-5 Scientific Review Committee. Psychological Medicine, 43, 1793–1800.Find this resource:
Nelson, H.L. (2001). Damaged Identities, Narrative Repair. Ithaca, NY: Cornell University Press.Find this resource:
Solomon, M. (2001). Social Empiricism. Cambridge, MA: MIT Press.Find this resource: