The Aftermath of the Concept of ‘Psychiatric Comorbidity’

Abstract

Accessible online at: www.karger.com/pps The concept of ‘psychiatric comorbidity’ (i.e., coexistence of two or more psychiatric diagnoses) has become very popular in the last few decades. ‘Psychiatric comorbidity’ has been found to be very common both in the general population and in clinical settings. For instance, in the US National Comorbidity Survey [1], only 26% of patients with a DSM-III-R/DSM-IV diagnosis of major depression had no concomitant (‘comorbid’) mental disorder, while in a study carried out in a general psychiatric clinic [2] more than one third of patients presenting for admission had at least three concomitant (‘comorbid’) DSM-IV axis I disorders. It has been repeatedly pointed out [e.g., 3, 4] that the term ‘comorbidity’ as originally defined by Feinstein [5] – i.e., the occurrence of a ‘distinct additional clinical entity’ during the clinical course of a patient having an index disease – should not be used to indicate the concomitance of two or more psychiatric diagnoses, because in most cases, due to our current very limited knowledge of the etiopathogenesis of mental disorders, it is unclear whether the concomitant psychiatric diagnoses actually reflect the presence of ‘distinct clinical entities’ or refer to multiple manifestations of a single clinical entity. ‘Disorders’ are different from ‘diseases’ [6], and even the term ‘disorders’ (rather than ‘syndromes’, i.e., constellations of symptoms) is probably inadequate to reflect the status of our current diagnostic categories [7]. Moreover, the emergence of the phenomenon of ‘psychiatric comorbidity’ has been to a large extent a by-product of some characteristics of our current diagnostic systems, such as ‘the rule laid down in the construction of DSM-III that the same symptoms could not appear in more than one disorder’ [8], the proliferation of diagnostic categories, the limited number of hierarchical rules, and the fact itself that the current systems are based on operational diagnostic criteria, which may be less able than traditional clinical descriptions to convey the ‘gestalt’ of some diagnostic entities [4]. The recent debate on ‘clinimetrics’ vs. psychometrics in this journal [9, 10] is enlightening in this respect. The main argument which has been put forward [e.g., 11] to respond to the above criticisms is that, exactly because our current knowledge of the etiopathogenesis of mental disorders is very limited, we should adopt an atheoretical and descriptive approach, allowing to record all the diagnoses whose criteria are fulfilled by each individual, with very few hierarchical rules. This would ensure the collection of a greater amount of clinical information, a more comprehensive and targeted approach to treatment and the exploration of the pathophysiological correlates of the individual ‘comorbid’ mental disorders. Is this argument convincing? Not completely, as its proponents themselves recognize [e.g., 11, 12]. The assumption that encouraging multiple diagnoses allows the collection of a greater amount of information in clinical practice remains at present not proven by empirical research. What research evidence actually suggests is that clinicians tend not to record all the diagnoses that a given individual fulfils [2, 13]. Especially in very busy practices