Comments on “Psychopathology in Neuropsychiatry: DSM and Beyond”

Vaidya and Taylor have made a valuable contribution to the literature on current neuropsychiatric practice and its reliance on the Diagnostic and Statistical Manual of Mental Disorders (DSM). In their article entitled “Psychopathology in Neuropsychiatry: DSM and Beyond,”¹ they are critical of DSM classifications and the straightjacket into which these classifications have thrust modern psychiatry. In particular, they expressed concern that descriptive psychopathology has become a thing of the past, seldom taught to trainees, and that neuropsychiatric diagnostic skills and practice have suffered as a result.

In their article, Vaidya and Taylor suggest that the popularity of DSM criteria results, at least in part, from the needs of psychopharmacologists and that a reliable diagnosis, in terms of DSM, is viewed as the first step in selecting from “an array of pharmacotherapy algorithms.” Yet reliability does not necessarily imply validity.

I cheered when I read the authors’ statements that:

In order to examine this issue further, Vaidya and Taylor surveyed all accredited psychiatry residency training programs in the U.S. by mail and inquired about the number of courses and teaching hours devoted to psychopathology, the use of “classical” textbooks on psychopathology, and “classic” topics of psychopathology covered in each training scheme. From a 45.6% response rate, they found that approximately three-quarters of the programs surveyed offered courses in descriptive psychopathology and in mental state assessment, but fewer than 50% of programs offered a course in the use of DSM. Only 33% taught trainees how to identify the Capgras and Fregoli syndromes, and, perhaps most strikingly, fewer than 50% offered instruction on how to distinguish phenomena such as delusional perception, delusional mood, secondary delusional ideas, and over-valued ideas, or how to elicit and identify experiences of alienation and control. Furthermore, fewer than 20% made use of Schneider’s Clinical Psychopathology; 6% used Jaspers’s General Psychopathology; and only one program employed Fish’s Schizophrenia.

While some of the individual symptoms that Vaidya and Taylor inquired about were rather idiosyncratic—and it may not be desirable to have a full course (rather than general guidance) in how to use DSM—their point was well made. It is particularly worrisome that one-quarter of the training programs surveyed did not appear to teach descriptive psychopathology at all, and it is reasonable to fear that this proportion is growing. (In the United Kingdom, the same trend is evident, exacerbated by the way that community psychiatry is currently being organized and managed within the National Health Service).

Vaidya and Taylor described two cases—presumably personal cases—in which only an understanding of descriptive psychopathology, but not DSM, would allow accurate diagnosis and management. I am not entirely convinced that these two cases are the best examples, and I had questions about the findings of an electroencephalogram in the second case. Nevertheless, the authors’ point that neuropsychiatry is poorly catered for in DSM is absolutely correct. They accurately note that DSM does not incorporate any adequate description of the variety of neuropsychiatric syndromes that occur in epilepsy. The same point can be made about DSM’s coverage of other major neuropsychiatric disorders, such as traumatic brain injury, alcohol-related neuropsychiatric syndromes, and specific memory and other cognitive disorders. Moreover, the authors’ point that “The DSM does not consider patterns of features, but rather the number of features” is absolutely spot on, as there is no emphasis on an understanding of the pattern or logic of symptoms in either the DSM or the International Classification of Diseases (ICD), which is, if anything, even worse.

Simon Fleminger and I² made a number of related observations in a paper criticizing the cookbook style of such classificatory systems and their special inadequacies with respect to neuropsychiatry. Similar to Vaidya and Taylor, we argued that: “Too often, the present versions appear to give committee-driven compromise statements, together with arbitrary counts of the number of symptoms required to fulfill a diagnostic criterion, and this often gives rise to the impression of a lack of real understanding of the respective disorders. Nowhere is this more true than in the description and definitions of neuropsychiatric disorders.”

We also argued for a simplified classification, in which, for example, vascular dementia would not be subclassified as acute onset, multi-infarct, subcortical, mixed cortical and subcortical, other, and unspecified, as occurs in the ICD. Such classifications give the impression that mixed patterns are the exception rather than the rule and that clinical diagnosis is more precise than is usually the case. In contrast, we supported the introduction of modern diagnostic classifications (e.g., dementia with Lewy bodies (DLB), frontotemporal dementia (FTD), and the DSM classification “personality change due to … a general medical condition,” specifying the particular medical disorder). We also supported a proposal from MR Trimble³ suggesting that DSM and ICD should have links to other classificatory systems, drawn up by experts (e.g., the International Classification of Epileptic Seizures and the widely agreed upon diagnostic criteria for Alzheimer’s disease, DLB, or FTD).

Such a proposal would have a number of advantages in that it would start to draw together the various classificatory systems, and those for specialized disorders tend to have been drawn up by experts in those topics, whereas the DSM/ICD classifications often appear to have been compromises agreed by committees not necessarily expert on a particular disorder.

In our article, we argued that both DSM and ICD should distinguish much more clearly between progressive degenerative dementias and nonprogressive brain damage. At the same time, there needs to be, for each diagnostic category, some way by which severity can be classified (e.g., a classification of mild/moderate/severe specific memory disorder rather than having amnesic syndrome in one box and mild cognitive disorder in another). We also argued for better integration of neuropsychiatric diagnoses within both DSM and ICD, in the case of alcohol-induced neuropsychiatric disorders, traumatic brain injury, sleep disorders, and amnesic syndromes, for example. We advocated that, ideally, dissociative disorders would be much better integrated with other neuropsychiatric syndromes, or at least the links should be clearly signposted. On the basis of these observations, we proposed some major headings and subheadings by which neuropsychiatric disorders might be organized within revised versions of DSM or ICD.

In summary, the case for upgrading the status and descriptive accuracy of neuropsychiatric syndromes within DSM and ICD is overwhelming. There is no need for these resources to remain checklists or cookbooks designed for the purpose of counting symptoms. They could, instead, be reorganized and restructured to reflect a genuine understanding of the syndromes described. In this, we neuropsychiatrists, who perhaps have greater confidence in the neurobiological validity of neuropsychiatric syndromes, could set an example to our colleagues in general psychiatry. Vaidya and Taylor are absolutely correct in their description of the DSM and ICD: such books are only guidelines. They can never substitute for a proper understanding of descriptive psychopathology, and of the history of how our present concepts evolved. We forget this at our peril.