Home Psychology Psychiatric Diagnosis Revisited: From DSM to Clinical Case Formulation
As discussed in the previous three sections, psychiatric diagnosis was strongly under fire in the 1960s and 1970s: academic researchers revealed that its reliability and validity was insufficient, critical scholars exposed the subtle yet dark power dynamics underlying the diagnostic process and questioned the medical status of the discipline, and societal changes challenged treatment practices overall. These challenges harmed the psychiatric establishment, which was then largely populated by psychoanalytic psychiatrists. However, American psychoanalysts, who were incidentally criticized for their mediocre conventionalism by their European colleague Jacques Lacan (1953-1954, 1954-1955), could not deal very well with the challenges they were facing (Decker 2013). The overall effect was a steep decline in their influence on mental health care.
Young researchers with a renewed interest in biological psychiatry, by contrast, found an opportunity in this crisis and suggested reorganizing psychiatry along different principles. Particularly influential in this context were a group of psychiatrists affiliated with the Department of Psychiatry at Washington University in St. Louis. Their alternative approach consisted of reaffirming the biomedical approach to psychiatric disorders, endorsing its scientific status by embracing statistical methods and making a plea for biological research. Fundamental to this movement were the works of John Feighner, Samuel Guze, Eli Robins, and George Winokur (Decker 1997, 2013; Strand 2011). In the 1960s and early 1970s these authors promoted the idea that modern psychiatry should adopt Emil Kraepelin’s descriptive approach, hence their identification as “neo-Kraepelinians” by Klerman (1978, p. 104). They condemned psychoanalysis, argued that only research with a strong focus on biology could hold any promise for psychiatry, and, more generally, aimed to refute all psychological approaches to mental suffering.
Kraepelin was attractive to them because of his classificatory focus and his assumption that mental diseases have biological causes. Like a Linnaean botanist, Kraepelin concentrated on the form and essence of cases of mental disease, as well as their various presentations, and how they change over time (Berrios and Hauser 1988; Compton and Guze 1995). By describing the symptoms presented, their onset, course, outcome, prevalence, as well as predisposing factors, familial illness history, and risk factors, he aimed to detect disease-specific biological abnormalities. Kraepelin assumed that psychiatric conditions are syndromes that make up illness entities with a biological basis and origin: “Judging from our experience in internal medicine it is a fair assumption that similar disease processes will produce identical symptom pictures, identical pathological anatomy, and identical etiology” (Kraepelin 1907, p. 117). At first Kraepelin (1907, p. 115) believed he would discover these biological abnormalities: “The principle requisite in the knowledge of mental diseases is an accurate definition of the separate disease process. In the solution of this problem one must have, on the one hand, knowledge of the physical changes in the cerebral cortex, and on the other of the mental symptoms associated with them.” However, as his work progressed he postponed this dream, opting to instead develop a differentiation between psychiatric conditions by means of their associated prognosis. In his later work, Kraepelin grew more cautious of the idea of categorical differences between different disorders. While a large part of his work initially built on the distinction between dementia praecox and manic- depressive insanity, he later acknowledged a certain overlap between these conditions and entertained the possibility that they in fact reflect a unitary form of psychosis (Berrios and Hauser 1988).
The first idea that the neo-Kraepelinian psychiatrists put forward was that psychiatric discourse should be more strictly attuned to biomedical thinking. This implies that “concepts, strategies, and jargon of general medicine are applied to psychiatric disorders: diagnosis, differential diagnosis, etiology, pathogenesis, treatment, natural history, epidemiology, complications, and so on” (Guze 1992, p. 4). As indicated above, critical scholars like Foucault and Szasz had questioned the biomedical status of psychiatry. Along this way, they not only became archenemies of the establishment, but also seem to have provoked a reactionary move away from psychosocial lines of reasoning. The neo- Kraepelinians, who made up half of the DSM-III task force (Decker 2013, p. 98), were explicit in their wish to attune psychiatry to other medical disciplines, such as neurology. They thus revived the anatomo- clinical ideal of the nineteenth-century alienists, who aimed to connect clinically observable symptoms with anatomical lesions. Robert Spitzer, in his turn, was more subtle in his position. In the DSM-III he advocated an atheoretical position, which I discuss in Chap. 3, but nevertheless adopted many ideas from his biomedically inspired collaborators. In a later interview he suggested that, with the help of the DSM-III, psychiatry “looked more like a medical specialty” (Spitzer 2000, in Strand 2011, p. 300, my italics). In fact this is an interesting remark, as it leaves us wondering whether, in his opinion, it actually is a truly biomedical specialty.
The second idea that the neo-Kraepelinian psychiatrists promoted is that any further examination of mental disorders should start from “the primary organ of psychiatric illness” (Compton and Guze 1995, p. 200), that is, the brain. Indeed, in their view mental disorders are above all rooted in disorder-specific biological factors, and these make up the real substance of which symptoms, complaints, and behaviors are the observable form. Hence Compton and Guze’s (1995, p. 200) statement: “the medical model is without a priori theory, but does consider brain mechanisms to be a priority.” In making such a claim the authors act as though formulating priorities for one’s study of psychopathology is atheoretical. Yet any choice for a particular perspective in the study of a particular phenomenon obviously reflects a theoretical preference. By prioritizing the brain, and leaving aside aspects like socio-cultural context and life events (e.g., Van Os et al. 2010), which later proved to have a serious impact on severe mental disorders, a particular theoretical choice comes to the fore.
The third idea that the neo-Kraepelinian psychiatrists put forward was that diagnosis in psychiatry comes down to detecting psychiatric conditions by classifying symptoms, complaints, and behaviors based on pre-established criteria (Feighner et al. 1972; Welner et al. 1974). This brought them to make the now famous statement, which is often taken for granted, that “classification is diagnosis” (Robins and Guze 1970, p. 983, my italics). In their view, diagnosis is not concerned with the patient qua person but with the disorder a person is carrying. The underlying idea is that disorders can be studied independently of the particularities of people’s lives and thus merely require correct classification in terms of established lists of possibly occurring conditions (Klerman 1978). For them the field of psychiatry is presumed to be made up of discrete disorders that can be distinguished from one another based on five criteria: (1) clinical descriptions of symptoms that usually connect to a disorder; (2) differential diagnostic criteria that distinguish a given disorder from other conditions; (3) biomedical laboratory studies; (4) followup studies on the course and outcome of a disorder; (5) family studies for detecting patterns of heredity (Robins and Guze 1970; Feighner et al. 1972). The first two features make up the directly observable side of a disorder, and should, if clearly described, lead to a reliable classification of patients. The latter three components aim at detecting differences at the level of etiology and underlying pathophysiological illness mechanisms that can further explain the differences between disorders. Obviously, in their interpretation of Kraepelin, the neo-Kraepelinians focused on his early work, and assumed that neurobiological invariants could be found at the basis of observable symptoms. Along this way, Kraepelin’s later evolutions, which questioned categorical distinctions between disorders, were neglected. As already indicated, Spitzer strongly believed in these classificatory principles and assumed that this approach would enhance diagnostic reliability (Spitzer and Fleiss 1974). He consequently used them as a basis for elaborating the DSM-III disorder categories.
In the first paper that actually introduced the format of criterion- based diagnosis, Feighner et al. (1972) came to describe 16 disorders that cohered reasonably well with the five basic principles that should be fulfilled for each discrete category. Among others, these included schizophrenia, hysteria, obsessional neurosis, primary affective disorders (depression and mania), as well the then controversial category of homosexuality. All were considered to be reasonably valid diagnostic categories. In 1974 R.A. Woodruff and colleagues reduced this list to 12, and in 1978 Spitzer, Endicot, and Robins extended it again to 25 by elaborating the differentiation between forms of affective disorder (see Table 2.3). Indeed in 1978, only two years before the publication of DSM-III, Spitzer and colleagues published their Research Diagnostic Criteria (RDC), which aimed at enhancing reliability by providing standardized sets of factors that had to be evaluated. They tested these operationally defined disorder categories in three small-scale studies, using judges that were highly familiar with the RDC. One of these studies, including 60 patients, used a test-retest design: in a timeframe of two days two psychiatrists evaluated the same patient using a standardized interview schedule. The authors presented their observed kappa values as a scientific achievement: “The results of these three studies indicate that the reliability of the RDC categories is very high” (Spitzer et al. 1978, p. 779, my italics). Remarkably, in contrast to their earlier study, Spitzer and colleagues didn’t define the norms upon which they based the evaluation of their kappa values and as a consequence it is unclear how they came to the conclusion that the reliability of these diagnostic categories is “very high.” This might indicate that Spitzer and his associates were actually very selective in how they used the kappa statistic. If they had used the norms Spitzer used with Fleiss (1974), they would have concluded that out of the 18 conditions tested, 2 have a high reliability; 12 have satisfactory reliability, and 4 have unacceptable reliability (see Table 2.3). Indeed, it seems Spitzer and associates changed their standards in terms of the conclusions they wanted to draw. Even based on the norms of Landis and Koch (1977) or those used by Clarke et al. (2013) the assertion of “very
Table 2.3 Kappa coefficients for research diagnostic criteria disorders adapted from Spitzer et al. (1978, p. 779), and three sets of interpretations of the kappa coefficient using different norms
PEO present episode only, LD lifetime diagnosis high” reliability is clearly an overestimation. Such biased use of evaluation standards sheds doubt on the scientific integrity of the DSM-III architects. Above all, they seem to have been marketing a biomedical approach in psychiatry, which they disguised by presenting quantitative output that they interpreted in a partial way. Moreover, given the fact that those who tested the RDC were also those who developed them (adding to that the small sample used in the study), drawing such firm conclusions is unjustifiable (Decker 2013).
Compared to the DSM-II (1968), which counted 182 disorders, the lists of presumably valid diagnoses presented by Feighner et al. (1972), Woodruff et al. (1974), and Spitzer et al. (1978) were much smaller. Their trimmed lists of 16, 12, and 25 categories might respectively indicate that an age of new austerity was needed. Yet, this is not what happened. Despite its scientific aspirations the DSM-III was far less parsimonious than any system before, in terms of the 265 illness categories it discerned. Nonetheless, it was presumed that these categories reflected discrete disorders. Spitzer clearly relaxed the requirements that the neo-Kraepelinians had formulated for valid diagnostic categories: concentrating on symptomatic description alone, the DSM-III set aside the principle that diagnostic categories had to be based on biomedical laboratory studies, follow-up studies, and patterns of heredity (Robins and Guze 1970; Feighner et al. 1972).
In this context, Spitzer seemed to be using a double standard: while he believed that in research settings the number of diagnostic categories had to be limited in order to avoid false-positive diagnosis, he also argued that in clinical contexts, an extensive list of disorders with flexible criteria was needed in order to avoid false-negative diagnosis (Decker 2013, pp. 176-177). By making this choice, it looks like he wanted to boost diagnostic reliability in scientific studies, thus promoting the credibility of his approach. In clinical practice, by contrast, diagnosticians using the DSM-III had and still have to make choices based on far more elaborate lists of disorders, which increases the possibility that different diagnosticians use different labels to name the same problem. More recent naturalistic diagnostic research illustrates this problem. For example, in a study focusing on conduct disorder Kirk and Hsieh (2004) created a vignette with a short description of a 12-year-old boy.
The vignette was attuned to the diagnostic criteria of conduct disorder, as defined in the DSM-IV. Starting from this short description of the boy, a sample of 1334 psychiatrists, psychologists, and social workers with broad experience working with young adults (with an average age of 20.7 years) were asked to give him a DSM diagnosis. The group differed substantially in their diagnoses: 29 DSM clinical disorders were proposed as the primary diagnosis and only 45.5 % actually diagnosed conduct disorder, thus demonstrating that consistency in diagnosis is weak. Further examination of the data set demonstrated that brief descriptors of the boy’s ethnic background influenced the diagnosis significantly and that diagnosticians gave significantly different diagnoses based on their professional specialty and years of experience (Pottick et al. 2007). More fundamentally, clinicians’ assumptions about the presence of an internal dysfunction strongly determined the diagnostic choices they made (Kirk and Hsieh 2009). This shows that the interpretation of mental symptoms is not as strongly based on the descriptive features that diagnostic criteria supposedly capture, as people often believe.
Moreover, as the number of categories increased substantially in the DSM-III, it became increasingly unlikely that the plethora of new disorder categories, all created in the workgroup discussions (Decker 2013), would ultimately prove to be valid. Nonetheless, they were avidly used in clinical practice. Irrespective of relatively minor changes, these categories are also included in the DSM-5. Nevertheless, the DSM-5 differs substantially from the DSM-III and DSM-IV in terms of its basic assumptions concerning the disorder categories. Whereas the DSM-III and DSM-IV presumed that disorder categories are discrete, this is no longer the case in DSM-5 (Regier et al. 2013). Given the frequently observed comorbidity between different DSM conditions and the fact that biomedical studies fail to support the idea that DSM disorders correspond with delineated biological conditions, the DSM-5 committee gave up on the idea of discrete disorders (Hyman 2010). It is quite paradoxical, however, that the DSM-5 nevertheless holds on to the diagnostic categories defined by its predecessors.
|< Prev||CONTENTS||Next >|