Paul Bebbington
INTRODUCTION
In this chapter, I will deal with the difficult problem of classifying a disorder that looks more like the expression of a continuum than a useful category. The way affective symptoms are distributed in the general population calls into serious question the utility of a medical classification, and certainly makes procedures of case definition and case finding very difficult. Nevertheless, researchers do rely on these procedures to establish the epidemiology of the disorder, and in the second part of the chapter I will pull together recent findings on the prevalence and distribution of unipolar depression.
CLASSIFICATION AND UNIPOLAR DEPRESSION
The idea of unipolar depression is primarily a medical one; that is, it involves a particular way of looking at psychological disturbance. This centres on the notion of a syndrome that is distinct from other psychiatric syndromes. Some of these can be relatively easily distinguished-for example, paranoid schizophrenia-while others are acknowledged to be related. The disorders that most resemble unipolar depression are other affective disorders, that is, conditions that are characterized centrally by mood disturbance. They cover a number of anxiety disorders, other depressive conditions, and bipolar mood disorder.
Bipolar disorder is identified by the presence of two sorts of episode in which the associated mood is either depressed or predominantly elated. It is distinct from unipolar disorder in a variety of ways (such as inheritance, course, and outcome), and the distinction is therefore almost certainly a useful one. However, depressive episodes in bipolar disorder cannot be distinguished symptomatically from those of unipolar depression. As perhaps half of all cases of bipolar disorder commence with a depressive episode, this means that unipolar depression is a tentative category-the disorder will be reclassified as bipolar in 5% of cases (Ramana & Bebbington, 1995).
Psychiatric disorders are classified in the hope that the classification can provide mutually exclusive categories to which cases can be allocated unambiguously (case identification). Categories of this type are the basis of the medical discipline of epidemiology, which is the study of the distribution of diseases (that is, medical classes) in the population. This has been a very powerful method for identifying candidate causal factors, and is thus of great interest to psychiatrists as well as to clinicians from other specialities.
Syndromes are the starting point of aetiological theories, and of other sorts of theory as well-theories of course and outcome, of treatment, and of pathology (Wing, 1978). There is no doubt that the medical approach to malfunction has been a very effective one, generating new knowledge quickly and efficiently by testing out theories of this type (Bebbington, 1998).
SYMPTOMS AND SYNDROMES
The first stage in the establishment of syndromes is the conceptualization of individual symptoms. Symptoms in psychiatry are formulations of aspects of human experience that are held to indicate abnormality. Examples include abnormally depressed mood, impaired concentration, loss of sexual interest, and persistent wakefulness early in the morning. They sometime conflate what is abnormal for the individual and what is abnormal for the population, but they can generally be defined in terms that are reliable. Signs (which are unreliable and rarely discriminating in psychiatry, and thus tend to be discounted somewhat) are the observable concomitants of such experiences, such as observed depressed mood, or behaviour that could be interpreted as a response to hallucinations. Different symptoms (and signs) often coexist in people who are psychologically disturbed, and this encourages the idea that they go together to form recognizable syndromes. The formulation of syndromes is the first stage in the disease approach to medical phenomena, as syndromes can be subjected to investigations that test the various types of theories described above.
While syndromes are essentially lists of qualifying symptoms and signs, individuals may be classed as having a syndrome while exhibiting only some of the constituent symptoms. Moreover, within a syndrome, there may be theoretical and empirical reasons for regarding some symptoms as having special significance. Other symptoms may be relatively nonspecific, occurring in several syndromes, but, even so, if they cluster in numbers with other symptoms, they may achieve a joint significance. This inequality between symptoms is seen in the syndrome of unipolar depression: depressed mood and anhedonia are usually taken as central, while other symptoms (such as fatigue or insomnia) have little significance on their own. This reflects serious problems with the raw material of human experience: it does not lend itself to the establishment of the desired mutually exclusive and jointly exhaustive categories.
In an ideal world, all the symptoms making up a syndrome would be discriminating, but this is far from true, and decisions about whether a given subject's symptom pattern can be classed as lying within a syndrome usually show an element of arbitrariness. The result is that two individuals may both be taken to suffer from unipolar depression despite exhibiting considerable symptomatic differences.
This is tied in with the idea of symptom severity: disorders may be regarded as severe either from the sheer number of symptoms, or because several symptoms are present in severe degree. In practice, disorders with large numbers of symptoms also tend to have a greater severity of individual symptoms.
COMPETING CLASSIFICATIONS
The indistinctness of psychiatric syndromes and of the rules for deciding whether individual disorders meet symptomatic criteria has major implications for attempts to operationalize psychiatric classifications. There are currently two systems that have wide acceptance, the Diagnostic and Statistical Manual (DSM) of the American Psychiatric Association (APA) and the World Health Organization (WHO)'s International Classification of Disease (ICD). In the early days, revision of classificatory schemata relied almost wholly on clinical reflection. However, since the classifications are set up primarily for scientific purposes, they should properly be modified in the light of empirical research that permits definitive statements about their utility. The standardized and operationalized classifications that are now in existence offer an opportunity for using research in this way.
Unfortunately, much of the pressure for change has continued to originate from clinical and political demands. Revisions have sometimes had the appearance of tinkering in order to capture some imagined essence of the disorders included (Birley, 1990). What looks like fine-tuning can nevertheless make considerable differences to whether individual cases meet criteria or not, and thus disproportionately affects the putative frequency of disorders. We should jettison classifications only on grounds of inadequate scientific utility and as seldom as possible, since too rapid revision defeats the objective of comparison. Like all such classifications, DSM and ICD are created by committees. The natural tendency to horse-trading between experts selected precisely because they are powerful and opinionated leads to an over-elaborate structure, an excess of allowable classes and subclasses, and complicated defining criteria. Thus, in DSM-IV-R (APA, 1994), there are potentially 14 categories to which depressed mood can be allocated, and in ICD-10 (WHO, 1992) there are 22. Greater utility would probably accrue from limiting the primary categories to three (bipolar disorder, unipolar depressive psychosis, and unipolar non-psychotic depression), and epidemiological research often uses these categories in any case. In Table 1.1, I have provided a comparison of the definitions of depressive disorder under DSM-IV (APA, 1994) and ICD-10 (WHO, 1992), slightly simplified. Over the years, there has been considerable convergence between the systems. However, the differences remain important. The categories are too close together for empirical studies to establish their relative validity, but far enough apart to cause discrepancies in identification. Relatively severe cases are likely to be classified as depressive disorder under both systems. However, milder disorders may be cases under one system, and not the other. This becomes important in epidemiological studies of depressive disorder in the general population because such studies usually report their results under one system or the other, and the degree of comparability is hard to quantify. Thus, the use of different classificatory systems is one barrier to comparison between studies: there are others.
It is of interest to see the effect of applying algorithms for the diagnostic categories defined by different systems to a common set of symptom data. The Schedules for Clinical Assessment in Neuropsychiatry (SCAN) (WHO, 1992) allows diagnosis under both DSM and ICD. In Table 1.2, I have illustrated the effect of applying ICD-10 and DSM-IV criteria to the data from the Derry Survey (McConnell et al., 2002) on the identification of cases of depressive episode (ICD) and depressive disorder (DSM). Of the 18 participants diagnosed as having a depressive condition by one classification, two-thirds were diagnosed by both. Five cases of depressive episode were not diagnosed as DSM depressive disorder, whereas only one case of depressive disorder was not diagnosed as ICD depressive episode. In contrast, DSM recognized many more cases of anxiety disorder. Fifteen of the cases defined by DSM were not classed as anxiety disorders by ICD, while only two classified by ICD were not so classed by DSM. Thus, the ICD criteria appear to be less stringent for depressive episode, while the reverse is true of anxiety. The results suggest that the difference between the two systems arises because of differing thresholds rather than because of wide differences in the symptom contents of the classes.
THE LIMITS OF CLASSIFICATION
As classification aspires to 'carve nature at the joints', the empirical relationships between psychiatric symptoms create special difficulties of their own. In particular, symptoms are related non-reflexively: thus, some symptoms are common and others are rare, and, in general, they are hierarchically related, rather than being associated in a random manner. Rare symptoms often predict the presence of common symptoms, but common symptoms do not predict rare symptoms. Deeply (that is, 'pathologically') depressed mood is commonly associated with more prevalent symptoms, such as tension or worry, while, in most instances, tension and worry are not associated with depressed mood (Sturt, 1981). Likewise, depressive delusions are almost invariably associated with depressed mood, whereas most people with depressed mood do not have delusions of any kind. The consequence is that the presence of the rarer, more 'powerful' symptoms indicates a case with many other symptoms as well, and therefore a case that is more symptomatically severe. It is because of this set of empirical relationships between symptoms that psychiatric syndromes are themselves largely arranged hierarchically. Thus, schizophrenia is very often accompanied by affective symptoms, although these are not officially part of the syndrome. Likewise, psychotic depression is not distinguished from non-psychotic depression by having a completely different set of symptoms, but by having extra, discriminating, symptoms, such as depressive delusions and hallucinations.
LEAKY CLASSES AND COMORBIDITY
The operational criteria set up to identify and distinguish so-called common mental disorders cut across the natural hierarchies existing between symptoms. The consequence is that many people who have one of these disorders also meet the criteria for one or more of the others. This comorbidity has generated much interest, and was even responsible for the name of one of the major US epidemiological surveys (the National Comorbidity Survey) (Kessler et al., 1994). Researchers, then, divide into two camps: those who think the comorbidity represents important relationships between well-validated disorders; and those who think it arises as an artefact of a classificatory system that is conceptually flawed and fails adequately to capture the nature of affective disturbance.
Thus, Kessler (2000) has defended the status of generalized anxiety disorder (GAD) as an independent condition, despite its high comorbidity, arguing that it does, for example, precede major depression, and also outlasts it. However, this would be expected if GAD represented a low threshold disorder that could transmute into a higher threshold disorder with the addition of a few symptoms. GAD and depression certainly share a common genetic diathesis (Mineka et al., 1998). The superimposition of major depression on a long-lasting minor depressive disturbance (dysthymia) has been called double depression (Keller et al., 1997). The comorbidity of anxiety and depression may arise because anxiety states can transform into depressive disorders with the addition of relatively few symptoms (Parker et al., 1997). Depression/anxiety is equally apparent in adolescents (Seligman & Ollendick, 1998), as is the link between dysthymia and major depression (Birmaher et al., 1996). The idea that there are several distinguishable affective disorders is, to some extent, self-perpetuating, as it prevents clinicians from seeking out the full range of symptoms that are reflected in comorbidity. I imagine that it will turn out to be much more useful to see these comorbidities as an indication of common underlying processes leading to, but not necessarily reaching, a common destination.
DEPRESSION AND THE THRESHOLD PROBLEM
Another important empirical aspect of affective disorders is the distribution of symptoms in
the general population. Many people have a few symptoms, while few people have many.
This means that decisions have to be made about the threshold below which no disorder
should be identified. People who have few symptoms may still be above this threshold if
some of their symptoms are particularly discriminating, but, in general, the threshold is
defined by the number of symptoms. There is always a tendency in medicine to move the
threshold down, particularly as a sizeable proportion of the people with mental symptoms
who are seen by primary-care physicians fall below the thresholds of DSM-IV or ICD-10.
Continues...
Excerpted from Mood Disorders by Michael J. Power Copyright © 2004 by Michael J. Power. Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.