به سمت معیارهای تجربی برای طبقه بندی اختلالات شبه جسمی

There is a major need for an empirical evaluation of classification criteria for somatoform disorders. The present study analyzes psychometric properties of the existing criteria for somatization disorder. The full sample consisted of 324 patients seeking help because of “psychosomatic problems.” Data from a subsample of carefully diagnosed patients with somatization syndrome (n=76) and a clinical comparison group (n=32) permitted the analysis of the discriminative power of items. Twenty-one somatic symptoms adopted from DSM-IV and ICD-10 criteria did not exhibit the necessary psychometric characteristics (item probability, item-total correlation, etc.). Thirty-two somatic symptoms showed a satisfactory psychometric performance. A cut-off of seven or more symptoms yielded the best discrimination between low and high disability. New criteria for somatization syndrome (“polysymptomatic somatoform disorder”) are proposed taking into account for the strong association of somatization and abnormal illness behavior.

There are some major points of criticism with regard to the current classification of somatoform disorders. Although defined as a residual category, the most patients with persistent somatization fall into the category of “undifferentiated somatoform disorder” [1]. This is due to the rather restrictive diagnostic criteria of somatization disorder. The recent development of the DSM-IV and the ICD-10 did not change this fact substantially 2 and 3. Moreover, ICD-10 and DSM-IV criteria yield substantially different classification results [2]. The diagnostic situation is extremely unsatisfactory when one considers that somatization constitutes a very frequent phenomenon not only in European primary care settings but worldwide 4 and 5. Due to this critique, Escobar and colleagues 6 and 7 proposed an abridged somatization disorder. Patients had to endorse at least four (male) to six (female) somatoform symptoms out of the list of 35 DSM-III-R somatization symptoms. On the positive side, these altered criteria reflect the health-care relevance of the syndrome [8]. However, the empirical bases of the SSI-4/6-criteria were not evaluated in patients with functional somatic symptoms, but rather in patients with schizophrenia [9]. Moreover, the SSI-4/6 criteria are based on the DSM-III symptom list, which was not empirically derived. Due to the fact that the SSI-4/6 criteria have been used mainly in epidemiological research, their validity for use with clinical samples still needs to be addressed. It is likely that the SSI-4/6 criteria are overinclusive, and therefore fail to differentiate somatizing patients from those with other disorders. Kroenke and colleagues 10 and 11 also introduced a diagnostic category for patients with multiple somatoform symptoms below the cut-off for somatization disorder. They proposed the label “multisomatoform disorder.” This concept differed considerably from other current concepts as it stressed current somatoform symptoms and neglected lifetime symptom occurrence. The investigators presented interesting data analyzing several aspects of validity of their approach [10]. Further critique on the diagnosis of somatization disorder arose as consequence of its focus on symptom counting, whereas psychological or psychophysiological processes were virtually ignored 12 and 13. This also holds true for the abridged diagnostic criteria of Escobar et al. [8] or Kroenke et al. [10]. Although classification systems usually try to describe mental disorders using a multidimensional approach, the current classification of somatization disorder fails to include aspects such as behavior, cognitive attribution, biological arousal, personality, etc. Kirmayer and Young [14] claimed to take into account the cultural meaning of symptoms in the development of classification criteria. The World Health Organization has launched a study assessing medically unexplained symptoms in different cultures [15]. Although they found some culture-specific symptoms, they also reported ten symptoms with a mean frequency of >30% in patients at clinics in all centers (India, Italy, Zimbabwe, Brazil, USA). In a recently published article, we proposed new diagnoses for patients with somatoform disorders in three subgroups [13]: patients characterized by multiple somatic symptoms and abnormal illness behavior (polysymptomatic somatoform disorder); patients with extreme health anxiety (formerly hypochondriasis); and patients with few, but highly disabling somatic symptoms (specific somatoform disorder). Current patient subgroups with circumscribed somatic symptoms could be included in the last category (pain subtype, conversion subtype, chronic fatigue syndrome, and perhaps even recurrent abdominal pain, fibromyalgia, irritable bowel syndrome, etc.). However, we did not propose empirically derived criteria. In other studies, we demonstrated that typical cognitive and biological features of somatization exist 16 and 17. Patients with somatization syndrome show highly focused attention to bodily processes, hold overexclusive beliefs about good health, express catastrophizing interpretations of minor bodily misperceptions, and communicate a self-concept of being weak and intolerant toward stress [16]. These cognitive features resemble Barsky and Wyshak’s concept of somatosensory amplification [18]. Symptoms of physiological overarousal may also be present [17]. Therefore, multiple somatic symptoms without organic/pathological reasons, abnormal illness behavior [19], and certain cognitive-perceptual processes maintaining the disorder should constitute the main features of polysymptomatic somatoform disorder. For the diagnosis of somatization disorder, the classification systems of the DSM-IV and ICD-10 follow the same principles: In addition to some other features, patients have to report a minimum amount of bodily symptoms from a symptom list. This means that the number of symptoms is added up to a total score. The empirical basis for the symptom lists, for the addition of symptoms, or for the numeric cut-off point, however, is thin. This study aims to test the psychometric properties of such a procedure. A psychometrically driven analysis of all items should aid in clarifying the empirical value of current classification procedures for somatization disorder. Statistically, the computation of total scores only makes sense when single items have a significant and comparable influence on this summation index. Each symptom should show an intermediate frequency in high-risk groups, and should yield a significant item-total correlation. Moreover, the symptom frequency should differ between patients with and without somatization syndromes. If symptoms are too rare, they do not fulfill these criteria despite the fact that a specific symptom may be very disabling and may obviously not be due to a physical condition. This implies that rare symptoms should be considered elsewhere for classification of patient complaints, but should not contribute to simple summation scores. Moreover, to avoid an underestimation of the clinical relevance and frequency of symptoms, symptoms are not assessed in epidemiological samples, but rather in a clinical high-risk group.