Transcript
New Directions in the Treatmentof Somatization Robert L. Woolf olk, PhD a , b, *, Lesley A. Allen, PhD c , Jade E. Tiu, BA a ,c a Department of Psychology, Rutgers University, 152 Frelinghuysen Road, Piscataway,NJ 08854, USA b Department of Psychology, Princeton University, Green Hall, 18 Turner Court,Princeton, NJ 08544, USA c Department of Psychiatry, Robert Wood Johnson Medical School-UMDNJ,671 Hoes Lane, Piscataway, NJ 08854, USA I n this article the authors present their model of treatment for somatizationdisorder (SD) and related syndromes. It begins with a brief history of soma-tization followed by a discussion of theory and research on medically unex-plained symptoms. Finally, it describes in some detail the authors’ psychosocialtreatment for medically unexplained symptoms [1], which uses methods from both cognitive behavioral therapy and experiential emotion-focused therapy.Somatization is among the most puzzling phenomena that health careworkers encounter. In somatization physical symptoms occur in the absenceof any identifiable bodily mechanism. The causes of somatization that healthcare workers are able to implicate are neither proximate nor somatic, seeming instead to be indirect and to be characterized most aptly as psychologic or cul-tural. Somatization seems to be universal, found in all present societies and inall past societies for which there are relevant records.At the dawn of the psychoanalytic era, Breuer and Freud [2] developed theconcept of ‘‘conversion,’’ a process whereby intrapsychic activity putatively brings about somatic symptoms. Although Freud later broke with Breuerand created the substantial edifice of psychoanalysis, his work on hysteria pro-vided a blueprint for and harbinger of later theoretical explanations of somati-zation. Here the ideas of early emotional trauma or intrapsychic conflict as thecause of physical symptoms began to take shape. This work also introduced thenotion of a physical symptom as an unconscious form of communication, asa device for securing secondary gain, or a means for avoiding emotionalpain. The notion of the transduction of psychologic conflict into bodily symp-toms was disseminated widely as psychoanalysis began to dominate psychiatry. *Corresponding author. E-mail address :
[email protected] (R.L. Woolfolk).0193-953X/07/$ – see front matter ª 2007 Elsevier Inc. All rights reserved.doi:10.1016/j.psc.2007.07.001 psych.theclinics.com Psychiatr Clin N Am 30 (2007) 621–644 PSYCHIATRIC CLINICS OF NORTH AMERICA Stekel [3] coined the term ‘‘somatization’’ ( somatisieren ) during the early 1920sand defined it as ‘‘the conversion of emotional states into physical symptoms.’’A landmark in the descriptive psychopathology of somatization was the sem-inal monograph of Paul Briquet [4], Triate ´ Clinique et The ´rapeutique de L’hyste ´rie .The current conception of SD derives directly from this paper. Briquet’s metic-ulous and exhaustive listing of the symptomatology of hysteria remains unsur-passed. Briquet, in fact, described three related syndromes: conversionphenomena, hysterical personality, and multiple chronic unexplained somaticsymptoms [5,6]. These three syndromes overlapped somewhat in symptomatol-ogy, and they often were observed to co-occur. Briquet’s perspicuous work wasrevived by Purtell, Robins, and Cohen [7] and was developed further by mem- bers of the illustrious Washington University department of psychiatry. Perleyand Guze [8] published a list of 57 symptoms, commonly reported by womendiagnosed as having hysteria, that were clustered in 10 different areas. Theseinvestigators were the first to suggest specific criteria for the diagnosis of hyste-ria: the presence of 25 symptoms from at least 9 of the 10 symptom areas [9].Later, this list of 57 symptoms was expanded to include 59 symptoms, and theterm ‘‘Briquet’s syndrome’’ was adopted [10]. The criteria for Briquet’s syn-drome subsequently were incorporated into the Feighner criteria [11], the pre-cursor to the symptom set that appeared in the third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III) [12]. In that volume the theoreti-cally neutral term ‘‘somatization’’ was preferred to the more traditional termi-nology. Although some of the traditional language remains in the fourthedition of the DSM (DSM-IV) (eg, ‘‘conversion disorder’’), the word ‘‘hysteria’’no longer appears [13]. The ninth edition of the World Health Organization’s(WHO) International Classification of Diseases (ICD-9) [14], a more cosmopolitannosology of somatic and mental disorders published a year earlier thanDSM-III, retains much of the earlier terminology, including ‘‘neurasthenia’’as well as ‘‘hysteria.’’ The ICD-10 [15] has shifted in the direction of theDSM, although without banishing all the classical vocabulary.SDisdiagnosed,accordingtotheDSM-IV[13],whenapatienthasatleastfourunexplained pain complaints(eg, intheback,chest, joints),two unexplainedgas-trointestinal complaints (eg, nausea, bloating), one unexplained sexual symptom(eg, sexual dysfunction, irregular menstruation), and one pseudoneurologicsymptom (eg, seizures, paralysis, numbness). For a symptom to be counted to-wardthediagnosisofSD,itmustbemedicallyunexplainedorsubstantiallyinex-cess of the associated medical pathology. Also, each symptom must prompt theseeking of medical care or interfere with the patient’s functioning.The differences in the diagnostic criteria of DSM-III, DSM-III-R, DSM-IV,and ICD-10, although subtle in some instances, may be responsible for someof the inconsistencies in the epidemiologic findings. Some epidemiologic re-search has suggested that SD is relatively rare. In the Epidemiological Catch-ment Area (ECA) study [16], the largest survey of SD (performed ina community sample of 20,000 people across five sites in the United States),the lifetime prevalence of DSM-III SD was 0.13%. Prior research had resulted 622 WOOLFOLK, ALLEN, & TIU in higher estimates, ranging from 0.4% to 2% of the population, for Perley-Guze’s hysteria [17,18]. Not surprisingly, SD seems to be more common inprimary care settings than in community populations. In the WHO Cross-National Study of Mental Disorders in Primary Care [19], which assessed5438 primary care patients at 15 centers in 14 countries, the prevalence of SD was 0.9% and 2.8%, as defined by DSM-III-R and ICD-10, respectively.Smaller studies conducted in primary care settings have estimated the preva-lence of DSM-III-R SD to range from 1.0% in a sample of 685 patients [20]to 5.0% in a sample of 222 patients [21].For a number of reasons, several authorities suggest the actual prevalence of SD may be substantially higher than the literature suggests. First, autobio-graphic memory of past psychiatric symptoms, including somatization symp-toms, is unreliable [22,23]. Individuals seem to forget (or at least fail toreport) previously reported symptoms that are no longer troublesome[22,23]. Given that the diagnosis of SD requires patients to describe both cur-rent and remitted symptoms, and the latter often are not recalled, these studiesprobably underestimate the true occurrence of SD. Second, the diagnosis of SDrequires that a physical examination and diagnostic tests be performed or thatmedical records be reviewed to determine the nature of each symptom. Suchextensive investigations of physical symptoms are too costly to incorporateinto large epidemiologic studies. The third argument for the underestimationof prevalence rates has to do with the conjecture that physicians are more likelyto make the somatization diagnosis than are nonphysician diagnosticians[24,25]. Presumably, physicians are better able to distinguish between a medi-cally sound explanation for a symptom and the patient’s ‘‘medical sounding’’explanation for that symptom. Because nonphysicians conducted assessmentswithout access to medical records in the ECA and WHO studies cited previ-ously, SD may have been underdiagnosed in those studies. At the one site inthe ECA study where physicians evaluated symptom reports to determinewhether symptoms were medically explained, the prevalence of SD was higherthan at the other sites [26].Patients diagnosed as having the most severe form of somatization, SD, have been shown to incur health care expenses that are nine times the United Statesaverage, and they consume disproportionate amounts of the time and energy of health care providers [27]. In addition to the extensive direct costs, SD createsenormous indirect costs to the economy in the form of lost work productivity.Individuals diagnosed as having SD report being bedridden for 2 to 7 days permonth [27,28]. SD is not only costly but also is difficult to treat successfully. Ina longitudinal study following patients who had SD who were receiving stan-dard medical care, only 31% recovered after 15 years [29]. Typically, patientswho have SD are dissatisfied with the medical services they receive and repeat-edly change physicians [30]. These ‘‘treatment-resistant’’ patients frustratehealth care providers with their frequent complaints and dissatisfaction withtreatment [30]. Not one controlled medication trial for SD has been published,to the authors’ knowledge. Anecdotal evidence suggests many patients 623 TREATMENT OF SOMATIZATION diagnosed as having SD refuse to take medication, and those who do so fre-quently report adverse medication side effects [31]. The story is much thesame with other polysymptomatic somatoform disorders [32]. As of this writ-ing, pharmacologic treatment has had minimal success in somatization. SUBTHRESHOLD SOMATIZATION Some investigators have encouraged broadening the somatization construct toinclude the many patients affected by unexplained symptoms not numerousenough to meet criteria for full SD. Both the ICD-10 and the DSM-IV includeresidual diagnostic categories for subthreshold somatization cases. In the DSM-IV, undifferentiated SD is a diagnosis characterized by one or more medicallyunexplained physical symptom(s) lasting for at least 6 months [13]. The ICD-10’s criteria for undifferentiated SD differ from the DSM-IV’s in that ICD-10requires multiple unexplained symptoms lasting for at least 6 months [15].The ICD-10 provides an additional category for subthreshold SD, somatoformautonomic dysfunction, for cases of three or more unexplained symptoms of autonomic arousal [15]. The authors are aware of no published data that estab-lish the validity of any of these three diagnostic categories.Two research teams have suggested categories for subthreshold somatizationother than those described in the DSM-IV and the ICD-10. Escobar and col-leagues [33] proposed the label ‘‘abridged somatization’’ be applied to men ex-periencing four or more unexplained physical symptoms or to womenexperiencing six or more unexplained physical symptoms. Kroenke and col-leagues [34] suggested the category of ‘‘multisomatoform disorder’’ to describemen or women currently experiencing at least three unexplained physicalsymptoms and reporting a 2-year history of somatization. Both these subthresh-old somatization categories seem to be significantly more prevalent than fullSD, described earlier. Abridged somatization has been observed in 4.4% of community samples [33] and in 16.6% to 22.0% of primary care samples[19,20,35]. The occurrence of multisomatoform disorder has been estimatedat 8.2% of primary care patients [34].The demographic characteristic most often associated with somatization isgender. In the ECA study, women were 10 times more likely than men tomeet criteria for SD [36]. Gender differences, although not as extreme, alsohave been found in most studies employing subthreshold somatization cate-gories, such as Escobar’s abridged somatization or Kroenke’s multisomatoformdisorder [34,37]. A more complex picture of the association between genderand somatization was suggested by the WHO’s Cross-National study [19] inwhich female primary care patients were more likely than their male counter-parts to meet the ICD-10 criteria for full SD but were no more likely to meetEscobar’s abridged somatization criteria. At least on the severe end of the con-tinuum, SD is uncommon in men. Gender differences are less obvious in thevarious subthreshold syndromes.Current thinking is that the low prevalence of SD in men may be explained,in part, by stereotypic male traits, such as a disinclination to admit discomfort 624 WOOLFOLK, ALLEN, & TIU and an unwillingness to seek medical treatment [38]. Also, physicians may beless likely to consider somatization as a possible explanation for a man’s symp-toms than for a woman’s symptoms [39]. Gender biases may cause physiciansto communicate with and treat male patients differently from the ways in whichthey communicate with and treat female patients. At this juncture, there areonly conjectural explanations for the different rates of somatization in menand women.Ethnicity, race, and education have been associated with SD and subthresh-old somatization. Epidemiologic research has shown somatization patients weremore likely to be female, nonwhite, and less educated than nonsomatizers[16,19]. Findings on ethnicity have been less consistent across studies. Inthe ECA study [16], Hispanics were no more likely than non-Hispanics tomeet criteria for SD. The WHO study [19], conducted in 14 different countries,revealed a higher incidence of somatization, as defined by either ICD-10 orEscobar’s abridged criteria, in Latin American countries than in the UnitedStates.Findings on the relationship between age and somatization have been some-what consistent. Most studies indicate that SD and abridged somatization aremore common in middle-aged and older patients (over 45 years of age) thanin younger patients [19,25]. Some research [16], however, has detected no association between SD and age. Children and adolescents, of course, areextremely unlikely to meet criteria for SD (and cannot, if one uses the DSM-IV criteria), perhaps because they have not lived long enough to acquireenough clinically significant somatization symptoms, especially sexual andmenstrual symptoms. CONCEPTUALIZATION OF SOMATIZATION A biopsychosocial conceptualization of somatization leads to specific psychoso-cial treatment strategies that include cognitive, experiential, interpersonal, and behavioral interventions. This model of somatization emphasizes the interac-tion of physiology, cognition, emotion, behavior, and environment (Fig. 1). Pa-tients presenting with somatization have higher levels of physiologic arousaland are less likely to habituate to a stressful task than control subjects [40].This physiologic arousal is compounded by a tendency to amplify somatosen-sory information; that is, these patients are hypersensitive to bodily sensationsthat are experienced as intense, noxious, and disturbing [41]. Further, somati-zation patients form negative cognitive appraisals of their physical sensations[42]. For example, they may believe that pain, fatigue, and/or discomfort of any kind are signs of disease. In addition to misinterpreting somatic sensations,some patients think catastrophically to the extent that they imagine persistentphysical sensations to be a sign of some potentially fatal disease, such as canceror AIDS.Dysfunctional cognitions may elicit negative emotions or be elicited by neg-ative emotions [43]. This cognition–emotion cycle may interact in a complexfashion with maladaptive behaviors. For example, thoughts of possible illness 625 TREATMENT OF SOMATIZATION
