Prototype diagnosis of psychiatric syndromes and the ICD‐11

Drew Westen presents an alternative approach to the diagnosis of psychiatric syndromes, based on prototype matching. First, the article thoroughly reviews the problems posed by the current polythetic or count/cutoff methods of psychiatric diagnostic procedure, which were derived from the Research Diagnostic Criteria of the 1970s. It then details the advantages of using the new prototype-matching approach to diagnosis, in which diagnosticians compare a patient’s overall clinical presentation to paragraph-length descriptions of empirically identified disorders, and rate the “goodness of fit”, or how well the patient’s clinical presentation matches these prototypes. This method has been developed and tested by Westen and colleagues, mostly in mental health settings. My comments will focus on the possibility of incorporating Westen’s prototype-matching approach into the revision of ICD-10 Chapter V (F): Mental and Behavioural Disorders, which is now underway. As Westen correctly points out in his article, the Clinical Descriptions and Diagnostic Guidelines for ICD-10 Mental and Behavioural Disorders (the clinician version) have important similarities to a prototype-matching procedure, because they present what are usually paragraph-length descriptions of the clinical features of each disorder. In my opinion, Westen’s proposed model may have some advantages: namely, that it allows for greater flexibility and is more directly comparable to how clinicians think about patients. Moreover, it could be useful in research, teaching and training. However, I do not believe that the model can be incorporated as such into the revised chapter on mental and behavioural disorders in the forthcoming ICD-11. In particular, its rating procedure is problematic for a classification system like the ICD, which aims to encompass every kind of medical and mental condition, and targets a wide variety of users around the world. The ongoing revision of the ICD-10 Chapter V (F): Mental and Behavioural Disorders is occurring within the context of the revision of the entire ICD-10. The overall revision process has established rules for presenting information and for coding the presence or absence of the different disorders, as well as uniform requirements for the description of every disorder within the entire system. An attempt to use a different system of description and scoring as the basis for the chapter on mental and behavioural disorders would be against the general rules of the classification system as a whole, and undermine the parity of psychopathology with the rest of the medical system for clinical, administrative, and financial purposes in health care. In addition, Westen’s proposed system, as presented, loses any apparent advantage in clinical utility if we consider that mental health professionals are not the only ones involved in the diagnosis and classification of mental disorders. In fact, only a very small percentage of individuals with mental disorders will ever see a psychiatrist or any other type of mental health professional. Therefore, psychiatrists, clinical psychologists and psychiatric nurses cannot be envisioned as the primary users and the sole professional constituency for the ICD classification system – many other professional groups will also be using the classification. This includes primary care physicians as well as lay health care workers who deliver the majority of primary and mental health care in some developing countries. Asking these professionals to differentiate between a score of “4) good match: patient has this disorder, diagnosis applies”; and “5) very good match, patient exemplifies this disorder, prototypical case” would likely create confusion and uncertainty and reduce the clinical utility of the system. Furthermore, asking them to consider scores such as “2) some match, patient has some features of this disorder”; and “3) moderate match, patient has significant features of this disorder” could unnecessarily prolong the diagnostic procedure and lead to inflation in the diagnosis of subthreshold conditions. Reed et al 1 recently highlighted that the World Health Organization is concerned about the proliferation of diagnoses of mental disorders. As the International Advisory Group to the Revision of ICD-10 Mental and Behavioural Disorders has pointed out 2, all decisions concerning changes in the current classification should consider whether the proposed changes provide an improved basis for efficiently identifying people with the greatest mental health needs when they come into contact with health care systems. Although subthreshold conditions are increasingly being recognized as an important topic for research, this does not automatically mean that they should be defined as a disease, or included in the diagnostic formulation to the extent proposed by Westen. The new ICD version needs to be simpler, and also needs to pay special attention to the differentiation of what is a disorder from what is not. The main challenge in developing the new classification of mental and behavioural disorders is to identify the relevant threshold that clearly signals the presence of a condition deserving clinical attention, and to establish differentiations among conditions that have clinical utility 3,4. Attempting to integrate Westen’s approach would likely sidetrack this objective. Some elements of the rationale described for the content of the prototypes in Westen’s article could be relevant to revising the descriptions presented in the ICD system. One of the key elements in prototype-matching is that it centers on the fact that what matters to the clinician is the “gist”: a set of salient symptoms which, when present, are “good enough” for the clinician to establish a diagnosis, without the need to check the presence of other symptoms that are less relevant to making a diagnosis. The clinical descriptions incorporated into the ICD could take into account this need to emphasize the conditions’ salient features, and give less weight to symptoms that are less relevant to determining a given diagnosis.