The paper by de Waal and colleagues (2004, this issue) reports on the prevalence of somatoform disorders in Dutch primary care. They found that at least one out of six patients seen by general practitioners could be regarded as having a somatoform disorder, almost all in the non-specific category of undifferentiated somatoform disorder. The prevalence of the condition has major implications for medical services but what does this diagnosis mean? Is receiving a diagnosis of somatoform disorder of any benefit to the patient? Does it help the doctor to provide treatment?

THE SOMATOFORM CONCEPT

The category of somatoform disorders was introduced into DMS–III (American Psychiatric Association, 1980) to accommodate those patients who had somatic symptoms unexplained by a medical condition but with too few psychological symptoms to merit an alternative psychiatric diagnosis. The ICD–10 followed suit, although embedding somatoform disorders within a broader category of neurotic, stress-related and somatoform disorders (World Health Organization, 1992).

Somatoform disorders include a hetero-geneous group of diagnoses united only by their tendency to present with somatic complaints. In DSM–III–R (American Psychiatric Association, 1987) the specific subcategories included somatisation disorder, hypochondriasis, body dysmorphic disorder, conversion disorder and chronic pain disorder, but the classification proved inadequate to the clinical task and the most recent edition of DSM (DSM–IV; American Psychiatric Association, 1994) added the non-specific category of undifferentiated somatoform disorder. This diagnosis, which amounts to little more than relabelling the patient's own complaint, has turned out in practice to be the most common of the somatoform diagnoses, as reported by de Waal et al (Reference de Waal, Arnold and Eekhof2004, this issue).

LIMITATIONS OF THE SOMATOFORM DISORDER CLASSIFICATION

Despite its value in drawing attention to hitherto neglected patients, many clinicians and researchers have found this classification to be unsatisfactory (Reference Escobar and GaraEscobar & Gara, 1999; Reference MartinMartin, 1999). Its limitations are both theoretical and practical.

The main theoretical limitation is the assumption of psychogenesis. Although intended to be aetiologically neutral, the somatoform disorder concept in fact implied that the cause of these symptoms was understood, and was a mental disorder. The cause of somatic symptoms in the absence of disease pathology is not well understood, and the evidence suggests that it is likely to be multi-factorial including physiological (or minor pathological) as well as psychological and social factors (Reference MayouMayou, 1991). The simplistic somatoform disorder concept also encourages us to think of patients as having either somatic symptoms that are unexplained by disease or somatic symptoms that are well explained by disease. In fact, even in patients with disease, the symptoms are often not well explained by the identifiable pathology. For example, the symptoms of patients with cardiovascular disease have only a limited relationship with objectively measured disease severity (Reference Ruo, Rumsfeld and HlatkyRuo et al, 2003). Hence, the use of the concept of ‘unexplained by a medical condition’ as the core feature of somatoform disorders is problematic.

Somatoform disorders also have a number of practical limitations.

1. (a) The main limitation is that the psychogenic implication of the diagnosis is simply unacceptable to many patients, making it a poor basis for collaborative management (Reference Stone, Wojcik and DurranceStone et al, 2002).

2. (b) The labelling of a somatic presentation as a somatoform disorder may lead to the underdiagnosis of depression and anxiety. De Waal and colleagues have highlighted just how common this combination of diagnosis is (2004, this issue).

3. (c) Somatoform disorders do not relate to the parallel and more widely used general medical classification of functional somatic syndromes (such as irritable bowel syndrome and chronic fatigue syndrome) that are used in primary care and general medicine to describe the same patients (Reference Wessely, Nimnuan and SharpeWessely et al, 1999).

4. (d) Neither the overall category of somatoform disorder nor its sub-categories satisfy the accepted criteria for validity or reliability (Reference Kendell and JablenskyKendell & Jablensky, 2003).

5. (e) The value of somatoform diagnoses in guiding treatment is limited and often taken simply to indicate a need to minimise access to medical care (Reference Smith, Monson and RaySmith et al, 1986).

The severity of these shortcomings means that mere tinkering is unlikely to be adequate and that more radical revision is required. The ambitious programme to prepare for the forthcoming DSM–V and ICD–11 (Reference Phillips, Greenberg, Rasmussen, Phillips, First and PincusPhillips et al, 2003) offers an opportunity to reconsider the somatoform disorders. We suggest that both the term and diagnostic category have proved to be unhelpful and, therefore, should be abolished.

WHAT SHOULD REPLACE THE SOMATOFORM DISORDERS?

If the somatoform disorders are to be abolished, with what should we replace them?

First, we should adopt a new terminology that both avoids questionable aetiological assumptions and is acceptable to patients. A variety of alternative terms have been suggested, including ‘medically unexplained symptoms’ and ‘functional somatic symptoms’. None is ideal. We might do best to use a simple description of the symptoms, as is done for chronic pain. An alternative is to use the term ‘functional symptoms’, which, in its original usage (Reference TrimbleTrimble, 1982), implied merely a disturbance of bodily functioning rather than structure and did not presume a simple psychogenesis. We have also found that the diagnosis of ‘functional symptoms’ is relatively acceptable to patients (Reference Stone, Wojcik and DurranceStone et al, 2002).

Second, many of the specific diagnoses currently within the category of somatoform disorders could readily be ‘rehoused’ elsewhere in the existing classification. For example, somatisation disorder is arguably better considered as a combination of a personality disorder and anxiety and depressive syndromes (Reference Bass and MurphyBass & Murphy, 1995). Hypochondrias could be rehoused comfortably within the anxiety disorders (health anxiety; Reference Warwick and SalkovskisWarwick & Salkovskis, 1990). Body dysmorphic disorder could be placed in anxiety or obsessive–compulsive disorder and conversion disorder could be reunited usefully with dissociative disorders. The remaining categories (undifferentiated somatoform disorder, pain disorder and somatisation disorder) are all currently defined solely in terms of number or type of somatic symptoms. They might, therefore, be better considered as medical conditions and coded on Axis III. Indeed, the DSM–V coding already allows Axis III to be used to list functional somatic syndromes or symptoms such as pain, and this practice could be formalised.

There is often the need for more than a simple diagnosis to characterise patients accurately for the purposes of treatment and research. The new classification could accommodate this knowledge by en encouracouraging the use of a supplementary multiaxial description that includes not only the type and number of symptoms, but also the relevant cognitions, behaviour and physiological disturbances (Reference Sharpe, Mayou, Bass, Mayou, Bass and SharpeSharpe et al, 1995). There are already analogies in the DSM–IV sections on pain and sleep in their reference to other more detailed specialist classifications.

THE BENEFITS OF ABOLISHING SOMATOFORM DISORDERS

It has been argued that the somatoform disorder category has been valuable in drawing attention to a neglected waste-land between the walled citadels of medicine and psychiatry (Reference Bass, Peveler and HouseBass et al, 2001). In arguing for a firmer epidemiological base for this diagnosis, de Waal et al (Reference de Waal, Arnold and Eekhof2004, this issue) appear to accept this view. In contrast, we believe that the diagnosis has been unhelpful by perpetuating dualism. Medicine can only continue to assume that all their patients’ illnesses are accounted for by disease pathology if they can label those patients whose somatic complaints do not fit this assumption as really ‘psychiatric’. Psychiatry can only continue in its belief that these somatic complaints are really based purely in psychopathology by labelling them as having somatoform disorders, with the associated and dubious implications of ‘somatisation’ (Reference DeGucht and FischlerDeGucht & Fischler, 2002).

We predict that the abolition of somatoform disorders would have a positive impact on both medicine and psychiatry. Medicine would have to accept that many of its patients are poorly served by a focus solely on pathologically defined disease, and pay more attention to symptoms. Psychiatry would have to acknowledge that many of their patients are not well served by a focus solely on supposed psychopathology, and place more weight on patients’ protestations that their symptoms are genuinely physical. Indeed, we believe that it would be beneficial for all medical practice if there were more emphasis on the understanding and management of patients’ symptoms in their own right, rather than as merely signposts to diagnosis (Reference Kroenke and HarrisKroenke & Harris, 2001). What would medical care without somatoform disorders look like?

HEALTH CARE WITHOUT SOMATOFORM DISORDERS

The size of the problem of symptoms in primary care has been highlighted by de Waal and colleagues (Reference de Waal, Arnold and Eekhof2004, this issue). Clearly, in any future health care system primary care must continue to have a central role in the management of patients’ symptoms. However, it seems to us to be naïve to assume naïve that the problem of somatic symptoms could be managed effectively by primary care alone. What might be the role for psychiatry? Existing mental health services seem unlikely to provide the solution. The limited specialist liaison psychiatry and health psychology services will undoubtedly continue to have an important role. However, we argue that they represent only a transitional and partial phase of service development. In the ‘post-somatoform’ world we envisage that there will be a renewed interest by all parts of medicine in an integrated approach to patients’ symptoms. Such a development will require that psychological assessment and intervention are fully integrated into medical care. Readers who regard this vision as over-ambitious might be surprised to see it described as normal medical practice over 100 years ago (Reference Sharpe and CarsonSharpe & Carson, 2001).

CONCLUSION

In conclusion, the increasing recognition of the size of the problem of somatic symptoms unassociated with disease is well illustrated by de Waal et al (Reference de Waal, Arnold and Eekhof2004, this issue). The recognition of the importance of symptoms as a subject of study in their own right is both overdue and welcome. The question we all now need to ask is: does diagnosing these somatic symptoms as somatoform disorders help or hinder us, not only in our efforts to understand and treat this neglected group of patients, but also in our overall approach to human illness?