Approaches to Psychiatric Classification

Approaches to Classification

Etiological Versus Descriptive

Historically, there have been two fundamental approaches to formulating systems of psychiatric classification: etiological and descriptive. Etiology-based classification systems organize categories around pathogenetic processes so that disorders cor-responding to a particular category share the same underlying cause. Because the etiological basis for most psychiatric condi-tions remains elusive, etiological classification systems tend to be based instead on a particular conceptualization of the process of mental disorders. Although such classifications may be heuris-tically useful to proponents of the particular conceptualization that forms the basis of the system, they are often considerably less useful for proponents of different etiological principles, which greatly limits their utility. For this reason, a descriptive approach to classification has proved to be of greater utility given our current understanding. The descriptive approach aims to es-chew particular etiological theories and instead relies on clinical descriptions of presenting symptoms. This approach, advanced by the work of the 19th century psychiatrist Emil Kraepelin, formed the basis for the system of classification of the Diagnostic and Statistical Manual of Mental Disorders, Third Edition (DSM-III) introduced in 1980. As a result, DSM-III and its successors, the Diagnostic and Statistical Manual of Mental Disorders, Third Edition, Revised (DSM-III-R) and DSM-IV, have proved to be useful in a variety of different settings and by psychiatrists of widely different backgrounds and conceptual orientations.

Syndrome Versus Symptom

Given that the manual lacks a specific etiological conceptualiza-tion, what is its organizing principle? The fundamental element is the syndrome, that is, a group or pattern of symptoms that appear together temporally in many individuals. It is assumed that these symptoms cluster together because they are associated in some clinically meaningful way, which perhaps may reflect a common etiological process, course, or treatment response Although it was hoped that the syndromes identified in the DSM represented relatively homogeneous subpopulations of patients, over the past 20 years since the publication of these definitions in DSM-III, the goal of discovering common etiolo-gies for each of the DSM-defined syndromes has remained elu-sive. Epidemiologic and clinical studies have shown extremely high rates of comorbidities among the disorders, undermining the hypothesis that the syndromes represent distinct etiologies. Furthermore, epidemiologic studies have shown a high degree of short-term diagnostic instability for many disorders. With regard to treatment, lack of specificity in treatment response is the rule rather than the exception. The efficacy of many psycho-tropic medications cuts across the DSM-defined categories. For example, the selective serotonin reuptake inhibitors (SSRIs) have been demonstrated to be efficacious in a wide variety of disor-ders from many different sections of the DSM, including major depressive disorder, panic disorder, obsessive–compulsive dis-order, dysthymic disorder, bulimia nervosa, social anxiety disor-der, post traumatic stress disorder generalized anxiety disorder, hypochondriasis, body dysmorphic disorder and borderline per-sonality disorder. Results of twin studies have also contradicted the DSM assumption that separate syndromes have a different underlying genetic basis. For example, twin studies have shown that generalized anxiety disorder and major depressive disorder may share the same genetic risk factors (Kendler, 1996), and evidence from molecular genetics research (Berrettini, 2000) indicates that three of the putative susceptibility loci associated with DSM-defined bipolar disorder also contribute to the risk of DSM-defined schizophrenia.

Given these clear limitations in the syndromal approach, it is important that users of the DSM resist the temptation to reify the DSM diagnostic categories as if they were actual dis-eases. They are best viewed as clinically useful constructs that are helpful in facilitating communication and record keeping and in selecting treatment. As more information about the causes of mental disorders become evident over the next decades, it is more than likely that the syndromal approach will be replaced by a classification system that is more reflective of the underlying eti-ology and pathophysiology.

Categorical Versus Dimensional

The diagnoses included in DSM-IV are defined categorically, that is, diagnostic criteria are provided that indicate whether an individual’s clinical presentation either meets or does not meet the diagnostic criteria for a particular disorder. This method of classification is similar to what is used in other fields in medicine, namely that a patient either has or does not have a particular diagnosis, like pneumonia, colon cancer, multiple sclerosis, and so on. This tendency to define illness in terms of categories is undoubtedly due to the fact that it is reflective of basic human thought processes, embodied by the use of nouns in everyday speech to indicate categories of “things’’ (e.g., chairs, tables, dogs, cats).

In principle, however, variation in the symptomatology can be represented by a set of dimensions rather than by mul-tiple categories. An example of this in medicine is blood pres-sure, which is measured along a continuum from low to high. (It only becomes categorical when we apply the label “hyperten-sion’’ to indicate that a patient has a significant elevated level of blood pressure that puts him or her at risk for developing serious illness.) While a categorical approach to classification has important heuristic appeal, it may not represent the true state of things. Implicit in the categorical approach is an as-sumption that mental disorders are discrete entities, separated from one another and from normality, either by recognizably distinct combinations of symptoms or by demonstrably dis-tinct etiologies. While this has been shown to be the case for a small number of conditions (e.g., Down syndrome, fragile X syndrome, phenylketonuria, Alzheimer’s disease, Huntington’s disease and Creutzfeldt–Jakob disease), there is little evidence supporting the applicability of this model for most other psy-chiatric symptoms. Indeed, in the last 20 years, the categori-cal approach has been increasingly questioned as evidence has accumulated that the so-called categorical disorders like major depressive disorder and anxiety disorders, and schizophrenia and bipolar disorder seem to merge imperceptibly both into one another and into normality (Kendler and Gardner, 1998) with no demonstrable natural boundaries.

Dimensional approaches do have some clear advan-tages. First of all, the commonly observed phenomena of ex-cessive comorbidity (i.e., an individual receiving multiple, simultaneous DSM diagnoses) is arguably a direct result of having a categorical system with more than 250 categories. A dimensional approach, which would characterize an indi-vidual’s psychopathology by indicating the extent of his or her psychiatric symptomatology across a number of dimensions, virtually eliminates apparent comorbidity. For example, con-sider an individual who presents with depression, anxiety and social avoidance. Using the DSM-IV categorical system, cri-teria might be met for three diagnoses (i.e., major depressive disorder, social phobia and generalized anxiety disorder), thus warranting a diagnosis of all three disorders on Axis I. A di-mensional approach may simply indicate that the person has “high scores’’ on the depression, anxiety and social avoidance dimensions. Another advantage of the dimensional approach is that it avoids setting particular thresholds for distinguishing between pathology and normality. Rather than categorically saying that an individual has major depressive disorder, a di-mensional approach might say that the person is high on the depression dimension.

There are a number of practical problems that potentially limit the utility of adopting a dimensional approach. First of all, clinicians are accustomed to thinking in terms of diagnostic categories, and the existing knowledge base about the presen-tation, etiology, epidemiology, course, prognosis and treatment is based on these categories. Furthermore, decisions about the management of individual patients (e.g., whether to treat and with what type of treatment) are also much easier to make if the patient is thought of as having a particular disorder (with its associated prognostic and treatment implications) rather than as a profile of scores across a series of dimensions. It should be noted that, at least for personality disorders, current dis-satisfaction with the categorical approach has led to proposals for research that might allow for the adoption of a dimensional approach to classifying personality functioning in DSM-V (First et al., 2002).