Psychiatric Comorbidity: More Than a Kuhnian Anomaly

Dr. Aragona's article in this issue of Philosophy, Psychiatry, & Psychology makes some important points regarding the relationship between comorbidity rates and the classification system currently used in psychiatry. Particularly persuasive is his claim that observed patterns of comorbidity are, in important respects, consequences of the structure of the classification system. I am not convinced, however, that comorbidity is best conceptualized as an artifact of the taxonic structure of the Diagnostic and Statistical Manual of Mental Disorders (DSM) and the International Classification of Diseases (ICD). Instead, I suggest that the use of dichotomies such as true versus artifactual should be used with caution when psychiatric classification is being discussed.

In contrast with the artifact hypothesis, I argue that the phenomena covered by the term "comorbidity" are real, expectable features of the psychiatric domain. They are not the phlogiston and luminiferous ether of psychiatry. I am skeptical of the notion that high comorbidity rates as a whole can be understood as Kuhnian anomalies, or that there is a crisis, or that after a period of revolutionary science a new paradigm would eliminate these 'anomalies.' In short, high rates of comorbidity in psychiatry will not simply disappear once scientists develop a better nosological paradigm. High levels of comorbidity can, however, be both lowered and provided a better, evidence-based conceptual framework.

Many Conditions, One Organ

In general medicine, comorbidity conventionally refers to the simultaneous occurrence of two independent disease processes or distinct clinical entities, such as heart disease and liver cancer (Feinstein, 1970; Jablensky 2004). The presence of one disease is supposed to affect the other in such a way that the co-occurrence of the two has implications for prognosis and treatment.

Comorbidity in psychiatry is a broader concept, referring to a variety of conditions that overlap in multiple ways. The different kinds of overlap are amenable to detailed empirical analysis (e. g., Klein and Riso 1993; Neale and Kendler 1995). Depressive reactions to primary disorders, attention deficit symptoms associated with mania, panic attacks in the context of severe borderline personality disorder, substance abuse disorders among psychopaths, histrionic dynamics combined with narcissistic personality traits, and emergent conditions such as schizoaffective disorder are all distinct types of comorbidity. The kinds of overlapping conditions just listed are sometimes termed psychiatric comorbidity. Also important is comorbidity between psychiatric disorders and general medical conditions as is studied in behavioral medicine and health psychology. The crisis, if there is one, is a crisis of comorbidities.

The heart disease and liver cancer type of comorbidity as independent organ systems being affected is ruled out a priori in psychiatric comorbidity proper. In psychiatry, "psychopathological disease processes" can all be expected to affect a single organ, namely the brain or, more likely, affect the different systems that are tightly integrated in the brain. The tight integration can make putative comorbid psychiatric disorders difficult to disentangle.

It is a widely held consensus that causes, not organ systems, are what matter in comorbidity. In this tradition, Dr. Aragona claims that in general medicine shared symptoms between two diseases (such as fever and fatigue) are not a problem because the diseases are differentiated with respect to etiology; in psychiatry, however, where syndromes are differentiated by a description of symptoms only, symptomatic overlap is a problem.

Is descriptive psychopathology, therefore, a root cause of the "crisis of comorbidity" and should it be abandoned in favor of etiological classification? Unfortunately, forcing psychiatric nosology to conform to an etiological paradigm is unlikely to resolve the problem of psychiatric comorbidity. As noted, the problem is one of distinguishing between the different etiological disease processes that are affecting tightly integrated systems in the brain. The brain as a whole is a multilevel, dynamic, and open system, which has an ecology (Zachar 2000). As argued by Kendler (2005) and Schaffner (2002, 2008), it has not thus far, and may never be fully possible to segregate causes and their associated psychiatric disorders into parsimonious etiological packages. In the ecology of neuroscience, causes and effects tend to merge in a continuous series of common pathways.

It would be foolish...