Beyond diagnosis? The title of this book may sound intriguing to the reader, especially as, at first glance, cognitive–behavioural psychotherapy and psychiatric classification appear to be quite well suited to each other as evidenced by not only research publications but also textbooks on therapeutic techniques mostly presented according to diagnostic categories.
Are there any limitations with the psychiatric model? How can we define a relationship between psychiatric diagnosis and case formulation? Does case formulation require a diagnosis? And what is really necessary to know when understanding a problem and making a treatment plan? To answer these questions, it seems appropriate to produce reasons and developmental conditions for case formulation as a clinical approach.
Traditionally, clinicians dealing with behavioural disorders, especially in the psychiatric setting, were mostly expected to define and organize their clinical work in terms of nosological categories.
The practical end result of this was mostly classification and medication. There was little room for psychotherapy, and novel approaches were usually not encouraged. When, in the 1950s, behaviour therapy (BT) arrived on the scene, there seemed little willingness on part of the psychiatric establishment to change this tradition (Eysenck, 1990).
What are the problems with psychiatric diagnosis?
Apart from being merely descriptive, there can be considerable overlap between categories. Despite considerable improvements and refinements over the last two decades (APA/DSM IV-TR, 2000), this is largely true to this day, as Turkat (1990) has pointed out in the case of personality disorders: For example, when clinicians were asked to sort the criteria for all disorders to the matching categories, they were only able to assign 66% of these correctly, indicating lack of validity and diagnostic overlap (Blashfield & Breen, 1989). They also found a tendency to assign presented problems to multiple categories to (six to eight in the case of personality disorders).
More recently, Sturmey (2009) has provided an excellent critique from the methodological point of view about the shortcomings of the psychodiagnostic approach according to the medical model. In the field of BT, the adoption of diagnostic criteria when selecting treatment procedures came about in the 60s, in tandem with the development of standardized techniques evaluated in randomized controlled trials (RCTs). Although this approach is presented in a very persuasive way by many professional bodies, mostly of a medical background, Sturmey points out the many limitations of this approach. Given the problems of validity and reliability of psychiatric diagnosis, it is questionable if homogenous groups for controlled trials can ever be constructed. Psychological processes in development are mostly ignored, and individualization of treatment is not encouraged.
Furthermore, any interaction between diagnosis and treatment is not contemplated in these designs as one treatment method is simply compared with another one, placebo or waiting list control, etc. Such a design does not allow us to determine whether a treatment effective for one diagnosis would also be effective for another diagnosis or whether it was the most effective treatment for a diagnosis. One might speculate that anxiety-based disorders associated with a range of symptoms might be best treated with a method addressing the mechanism of the disorder rather than focusing on individual complaints (symptoms), which are presumably the main focus of a diagnosis.
From a clinical standpoint, which deals with a unique patient, average improvement, as established in RCTs, does not make much sense – the average patient simply does not exist. Besides, significance in trials can be achieved with large size samples and sensitive measures. And usually, substantial number of individuals improve little or not at all. And even with those who do improve, the effective ingredient remains unclear and might include non-specific effects such as the therapeutic interaction.
Taking these points together, it seems highly questionable whether parametric statistical methods are appropriate in evaluating the efficacy of psychotherapeutic treatments. On the assumption that treatment should be individualized, single-case experimental design methodology might be used instead.
Sturmey (2009) points out further limitations of the diagnostic model: When medication is recommended, individual responses can be varied and often quite problematic in the long term. For example, if perceived effectiveness of treatment is seen to lie outside one’s own resources, dependency on medication may occur and self-efficacy arising out of personal coping responses is not encouraged.
Standardized psychological treatments are equally problematic as the response to them is unpredictable, improvements are often not maintained or some problems remain unchanged.
For all these reasons, creative and experienced clinicians tend to reject the straightjacket of the diagnostic model and prefer an individualized approach.
Finally, reliability and validity issues are a continuous problem with ongoing DSM and ICD (World Health Organization, 2010) revisions. Research trials on these criteria are few and far between and are usually conducted after criteria have already been established. Sturmey also points out that clinicians tend to use their own methods and not the recommended structured assessment procedures that may achieve better reliability.
Unfortunately, over time, these developments did lead to an increasing medicalization in treating behavioural disorders – a far cry from what was originally intended by the pioneers of BT such as Eysenck, who preferred to focus on a learning principle-based analysis of individual cases. This was motivated by complex and difficult cases that were rejected by the established providers, usually psychiatry or dynamic psychotherapy (Eysenck, 1990).
As the latest revision towards DSM-5 clearly demonstrates, scientists and clinical experts still share many disagreements and, despite considerable non-medical input, there does not appear much progress regarding advice and guidance for non-medical treatments.
It is difficult to see how, after many controversial revisions, DSM-5 will ever deliver a classification system acceptable to clinicians and sufferers. The recent response by the BPS (2011) addresses these expected shortcomings:
Another problem with diagnosis is the lack of interest in understanding the underlying psychological mechanism of a disorder. Obviously, this is quite unacceptable from a behavioural-learning perspective. Grouping problems into categories cannot advance this search for causes as learning biographies are ignored. Thus, psychiatric diagnosis will not help us to explain onset, development and maintenance of a behavioural disorder. Such information, however, must be considered to be crucial for a comprehensive case formulation with explanatory power.
Finally, psychiatric classification is clearly linked with concepts of mental illness and normality. Such labels are frequently experienced as stigmatizing value judgements by clients and have been shown to be counterproductive for learning-oriented therapy (e.g. Szasz, 1961). Despite great efforts towards operationally defined categories in diagnosis, the issue remains contentious, especially for non-medical psychotherapists.
However, it seems surprising that only recently claims were made to the contrary, i.e. the suggestion that medicalization of behavioural problems might provide relief for the sufferer (Markowitz & Swartz, 2007). There appears no substantial clinical evidence for this; however, labelling may be preferred by some individuals as it allows evading personal responsibility in preference for medicated treatment over psychological therapy. I believe this to be a short cut likely to undermine a personal locus of control in the development of new resources and coping behaviours, thus preventing enhancement of self-efficacy processes (Bandura, 1977).
These comments on the recent development of psychiatric classification clearly demonstrate that not much has been achieved since the publication of the first edition of this classification. There are no signs for a ‘bottom–up’ approach based on information experienced by problem sufferers and conceptualized according to case formulation methodology. Nevertheless, the commentary report of the BPS on the development of DSM-5 must be welcomed as a reminder that new initiatives are needed.
Although mostly inspired by psychologists, early developments of BT usually took place in psychiatric settings. However, right from the start, there was a difference of opinion how this approach should be applied to behavioural disorders. While psychologists preferred an experimental learning–based approach, psychiatrists saw the ‘symptom focus’ of this new method compatible with the medical model and were hopeful of finding symptom-focused techniques. It is also noteworthy that clinical psychologists in those early days were working under psychiatrists and were given limited scope to conduct BT without medical supervision. Originally, this method was expected by psychiatrist to supplement, like psychometry, psychiatrists diagnosis! (Eysenck, 1990).
In view of this background, it is hardly surprising that, apart from the pioneering experimental work carried out mostly by psychologists, efforts to develop treatment methods were targeted on technologies designed to match a diagnosis. These efforts were motivated by a need to satisfy the requirements of psychotherapy outcome studies, rather than addressing the real needs of individual patients. More recently, advances in ‘operationalized’ diagnosis (e.g. DSM-IV-TR, 2000) have facilitated the development of more disorder-focused treatment manuals. These tend to be more sophisticated as most are developed in clinical settings (see Wilson, 1996, 1997 for a discussion). Despite such substantial improvements, there remain strong doubts as to whether or not treatment manuals will ever be able to address fully the complexity of individual problems (Hickling & Blanchard, 1997; Malatesta, 1995a, 1995b). That is not to say that manuals, as a product of technical expertise, may not be useful in clinical training and therapy (e.g. the acquisition of technical skills by trainees) provided adjustments to individual cases can be made (see discussion later in this text).
To conclude, it appears obvious that psychiatric diagnosis has not become a facilitating tool for the development of individual-focused BT. Considering ongoing problems with reliability, validity, prognostic value and co-morbidity, there must be considerable doubts (e.g. BPS, 2011) that any psychiatric classification system does fulfil such criteria in a satisfactory way.
Despite such concerns, psychiatric diagnosis continues to be a powerful and deeply rooted system in mental health services across the world that casts a long shadow over the provision of psychological psychotherapies. And there is another, slightly more mundane, reason why, especially in the United States, the home of DSM, psychiatric classification is not indispensible: Usually, clinicians cannot expect remuneration by health insurance companies unless an established psychiatric diagnosis is submitted, together with matching techniques (I. D. Turkat, personal communication, 2012)!
Given the diverse natures of case formulation and psychiatric diagnosis, it may seem contradictory that early attempts to experiment with applications of learning theory took place in psychiatric settings. However, this is easily explained, as these settings were usually the location for complex and difficult-to-understand disorders, not suited for psychiatric intervention and even less so for psychodynamic approaches that tend to be highly selective. It was mostly for these reasons that psychologists were encouraged to experiment with alternative approaches (Eysenck, 1990).
