| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
SOMATIC PRESENTATIONS: EPIDEMIOLOGY |
From the Department of Clinical Psychology and Psychotherapy (W.R.), Philipps University of Marburg, Marburg, Germany; and the Department of Clinical Psychiatry (G.R.), University of Chile, Santiago, Chile.
Address correspondence and reprint requests to Winfried Rief, Clinical Psychology und Psychotherapy, Philipps University of Marburg, Gutenbergstrasse18, D–35032 Marburg, Germany. E-mail: rief{at}staff.uni-marburg.de
ABSTRACT
Objective: To investigate the stability of somatoform symptoms/disorders.
Methods: A literature search was done to select studies reporting data on stability of medically unexplained physical symptoms.
Results: Whereas individual symptoms vary over time, grouping symptoms into syndromes seems to create stable features. There are substantial problems with the use of lifetime diagnosis, favoring classification approaches that require only present state symptoms. Further, doctors’ ratings that symptoms are "medically unexplained" is highly problematic and reduces interrater reliability. Misdiagnoses and overlooking of organic conditions are in the same range as for other psychiatric (and many organic) disorders; therefore, this does not seem to reduce the stability of the diagnoses of somatoform disorders.
Conclusions: These results indicate how the classification of somatoform disorders can be improved. Some new diagnostic criteria are suggested that could be considered in the revision of Diagnostic and Statistical Manual of Mental Disorders—V (DSM-V).
Key Words: somatoform disorder • classification • stability • reliability • misdiagnosis
Abbreviations: DSM = Diagnostic and Statistical Manual of Mental Disorders; TR = Text Revision; ICD = International Classification of Disease; GP = general practitioner
INTRODUCTION
The stability of a diagnosis is suggested as one of the major criteria for validity (1,2). Patients with one diagnosis should not fall into the category of another diagnosis at follow-up assessment. Robins (2) felt that accurate diagnoses should show that the disorder predicts course and does not predict different disorders over time. Diagnostic stability is desirable to buttress validity, although it should be kept in mind that some syndromes are not stable and vary by nature (e.g., manic episodes, transient ischemic attacks); yet, nobody questions the validity of these episodic syndromes. Therefore, stability of a syndrome should not be considered as sine qua non for defining classification criteria. However, information on stability of symptoms and syndromes is important, as it can point to weaknesses of some criteria. Low stability of diagnoses can be caused by low reliability of some criteria, although other criteria are reliable and should be maintained in future classification attempts. Therefore, stability results can offer guidelines on how to improve classification rules.
In this paper, we will summarize studies presenting data on the stability of criteria for the classification of somatoform syndromes. Unfortunately, the large epidemiological surveys did not address the question of stability of somatoform complaints adequately (e.g., Epidemiological Catchment Area study) (3), or even disregarded these frequent health problems (National Comorbidity Survey) (4); thus, the empirical basis for our analyses is limited to smaller longitudinal studies. We will focus on studies providing data on the classification of multiple somatic symptoms. The discussion of validity criteria of other somatoform disorders like hypochondriasis can be found elsewhere (5–7). From all subgroups of somatoform disorders, those patients with multiple somatic complaints seem to be the most costly subgroup (8–10). Diagnostic and Statistical Manual of Mental Disorders—IV—Text Revision (DSM-IV-TR) offers the diagnosis of somatization disorder for patients with multiple somatic complaints. The diagnosis of somatization disorder has been criticized frequently (11,12), because this diagnosis does not apply for most patients with multiple somatic complaints. Beside the overexclusiveness of this category, we will also outline additional factors that reduce the stability of the diagnostic result.
Overview of Studies Investigating Stability of Criteria of Somatoform Disorders
Table 1 presents an overview of studies investigating stability aspects of the classification of multiple somatoform symptoms. Study selection resulted from a PubMed/Medline search (search items "somatoform" and "stability") as well as from individual literature search. It is obvious that many studies confirm the stability of the syndrome, whereas some studies found unexpected changes in symptoms or diagnoses. We highlight two extremes. 1) The study of Kent et al. (13) investigated 38 patients with somatization disorder per interview and chart review. Four years later, interviewers blind for the former diagnoses were able to confirm the diagnosis in 37 (97%) of the 38 patients. 2) Arnold et al. found that, even for patients with general somatoform disorders (not only somatization disorder), the syndromes were stable (14). In contrast to these encouraging results, a study of the World Heath Organization including 15 study sites in 14 countries found 74 patients with somatization disorder at first assessment, but confirmed this diagnosis only for 21 (28%) 12 months later, whereas 49 patients were newly diagnosed as somatization disorder at follow-up (15). The divergence of the results of the studies mentioned in Table 1 will be used in this article to analyze sources of unreliability.
|
Lifetime Versus Present State Diagnosis
Somatization disorder according to DSM-IV-TR (16) requires "a history of many physical complaints beginning before the age of 30 years ... ". Thus, somatization disorder is a lifetime disorder, not a present state diagnosis. This has several implications. A key question is whether people can ever be cured of this disorder, as the history of symptoms always remains. Moreover, this diagnosis requires that patients remember symptoms correctly over many years. Finally, remembering the onset of symptoms if the onset was months or years ago is definitely a source of erroneous reports (17).
The difference between lifetime symptom report and present state symptoms in somatoform disorders is impressive. Simon and Gureje (15) found that 61% of medically unexplained lifetime symptoms were not reported 1 year later. Although the prevalence rates of current somatization syndromes were comparable between the two assessment points, many people were not able to remember the symptoms they reported 1 year ago. However, it is unclear whether inaccurate recall is restricted to somatoform symptoms, or whether this holds true for the lifetime report of symptoms in general. Another study (18) demonstrated that the report of physical complaints is erroneous in general, not only in medically unexplained symptoms. In an 11-year follow-up study, the authors found that lifetime symptom report is unreliable, both for medically unexplained and medically explained symptoms.
If classification depends on memory of past symptoms, typical memory errors occur. Symptom report of past symptoms is highly dependent on current symptoms and current symptom intensity. People seem to remember former physical complaints better if they (still) suffer from these symptoms at present. Furthermore, if symptoms are currently very intense, their course is remembered as being longer and their duration is overestimated. An experimental investigation has shown that unpleasant medical procedures are remembered as less unpleasant if they are followed by a more positive procedure compared with no subsequent procedure (19). Thus, current body sensations including severity and how recently they have occurred may influence the accuracy of symptom recall.
Culture-Specific Aspects Influencing Stability
Another problem when using long time frames for classification is the fact that the use of terms for time periods is partially culture-bound. Although number of years, day of birth, etc., are important and frequently used anchors of time ratings in most developed countries, some languages of developing countries do not even have terms for time periods such as month or year, and do not register birthdays. If people do not know their birthday or their age, they cannot remember whether symptoms were present before the age of 30 years, as requested for somatization disorder according to DSM-IV; thus, this time criteria opens the door for cultural differences in the validity of diagnostic classification.
Another source of culture-bound instability of the diagnosis of somatoform disorders is the list of symptoms that are included. The frequency of individual symptoms varies substantially between cultures (20,21), although grouping of symptoms might increase stability and also might reduce cultural differences. The only study of Table 1 including not only developed but also developing countries was the one demonstrating low stability of somatization disorder (15). The other studies investigated samples from US, The Netherlands, Scandinavia, and Germany. Therefore, there is a substantial need to investigate more systematically cultural issues in future studies.
Stability of Somatoform Disorders in Younger People
Most mental disorders develop in youth and adolescence. Therefore, the longitudinal investigation of symptoms and syndromes in this age group challenges the aspect of stability the most, as symptoms are not yet chronic and symptom recovery is more likely during adolescence than it is later. A very sound epidemiological survey investigating psychiatric symptoms by means of structured clinical interviews was done in Germany (22). The authors assessed anxiety, depressive, and somatoform symptoms in 2548 adolescents and young adults. In this 4-year follow-up study, the authors found that somatization syndrome was as stable as depression, and more stable than anxiety disorder. About half of the young people describing depression or multiple somatic complaints confirmed these syndromes 4 years later. Thus, although there is some variation, the present state diagnosis of somatization syndrome seems to be in the stability range of other psychiatric disorders.
Minimum Duration of 6 Months
For the diagnosis of undifferentiated somatoform disorders or somatoform pain disorder, a minimum symptom duration of 6 months is required. Although the use of this time period leads to culture-dependent variations of stability, it seems to have stability-increasing effects in many countries. When all currently presented symptoms in primary care are considered, many symptoms improve or disappear in subsequent months or years (23). However, if only patients with a minimum symptom duration of 6 months are included, somatoform symptoms are more likely to persist until follow-up (14). Therefore, it seems to make sense to distinguish spontaneously resolving symptoms from more persistent bodily complaints. However, it should be kept in mind that primary care doctors also need terms or diagnoses to describe patients requesting investigations because of nonpersistent physical complaints (24); such a diagnosis, however, is unstable per definition.
Stability of Single Symptoms Versus Syndromes
Many patients have various coexisting somatic symptoms ("somatoform syndromes") instead of single symptoms. If syndromes instead of single symptoms are considered, the stability is much higher. For the discussion of this issue, two types of stability must be distinguished: a) the stability of the symptom report of individual patients, and b) the stability of overall prevalence rates in large samples. Even the provocative results of the Simon and Gureje trial demonstrated that the present state prevalence rate of somatoform syndromes was the same at first assessment and 1 year later, although the individuals fulfilling the criteria were different at the two assessment points. Leiknes et al. (18) investigated stability within individual patients and reported that the grouping of symptoms improved substantially the stability of symptom report. Using the Composite International Diagnostic Interview, the researchers found that between 22% and 100% of individual symptoms were lost to recall at follow-up 11 years later; this rate was substantially lower for groups of symptoms. Interestingly, the rate of forgetting symptoms was influenced by age and gender, with men tending to forget more symptoms, and younger respondents remembering slightly better at follow-up.
The increase of stability when grouping symptoms with a syndrome of polysymptomatic somatoform disorder has been shown impressively in another study investigating primary care patients (23). In a 5-year follow-up study of 500 patients, the authors showed that individual symptoms have high rates of resolution in the upcoming months and years. However, if patients had multisomatoform disorder, they were less likely to improve than patients with fewer medically unexplained symptoms. Sixty-seven percent of patients who had 
2 bothersome symptoms at baseline still had bothersome symptoms 5 years later. This rate increased to 94% of patients with bothersome symptoms 5 years later, if they had 5 bothersome symptoms at baseline. These results are confirmed by a recently published study (25) showing that 86% of patients with multiple somatoform symptoms (at least three symptoms) were reclassified correctly 1 year later.
To summarize, there is evidence that grouping of individual symptoms with syndromes increases the stability at follow-up. Probably even other diagnoses that are based on symptom counts (e.g., panic attacks, major depression) may not have stability of individual symptoms, but patients could still meet the criteria for the disorder although specific symptoms have changed over time.
Doctors’ Rating of Origin of Symptoms; Self-Rating Versus Expert Rating of Symptoms
The diagnosis of somatization disorder requires that physicians rate the origin of the physical complaints as "not fully explained by a known general medical condition." Therefore, physicians’ ratings about the origin of the symptoms are crucial for the diagnosis. However, this seems to be a substantial source of instability, as doctors vary tremendously in their rating systems. Fink, Rosendal, and Olesen (7) have shown that 37 general practitioners (GP) rated the prevalence of medically unexplained symptoms in their patients very differently (from 4% to 33%). However, this huge variation of GP ratings did not reflect real prevalence differences between medical offices, as the somatization scale of the Symptom Check List indicated a constant rate of about 30% of patients with increased somatization scores in all GP offices. Therefore, it can be postulated that doctors’ ratings about the origin of physical complaints is a substantial source of unreliability. Considering cultural issues, doctor’s assumptions about etiology lead to further interrater differences. In developing countries, doctors’ possibilities for expensive medical examinations are limited. They must decide about the origin of the symptoms with much less diagnostic information than a doctor who has access to multiple, highly sophisticated, and expensive examination results. As long as the diagnosis of somatoform disorder primarily depends on excluded organic conditions, this might be a further reason for cultural variations.
Despite this fact, expert ratings on the number of somatoform symptoms correlate substantially with self-ratings of patients (r = .73) (26). However, in the above study, experts used a structured and standardized interview to diagnose somatoform symptoms and were familiar with the medical charts. The study aimed to analyze whether expert ratings and patient ratings about the number of somatoform symptoms have different reliability. We reanalyzed the data of a sample described in a recently published study (27), including 295 primary care patients. The 6 months retest correlation for the number of somatoform symptoms is in the range of r = .72 (somatoform symptom count according to expert interview) and r = .75 (according to self-rating), which can be considered as substantially high. Thus, the variable "somatoform symptom count" reveals stable scores in both self-ratings and in expert ratings.
Stability of Associated Disability
It is not only the syndrome that seems to be stable, but also the associated physical functioning and quality of life. This was shown in a study by Dickinson and others (28) using the self-rating scale SF-36. Patients with multiple somatoform symptoms were in the lowest quartile of physical functioning, at first assessment and 1 and 2 years later. Thus, multiple somatoform symptoms are associated with persistent reduction in quality of life and continuing disability.
Are Somatoform Disorders Misdiagnosed Organic Conditions?
The detection of a medical disease that explains all the somatic complaints could be one reason for reduced stability of the diagnosis of somatoform disorders. The scientific community was alarmed after Slater published that many patients with "conversion or hysteria" had misdiagnosed serious, sometimes even life-threatening diseases (29). However, a reanalysis of all trials on misdiagnosed patients initially diagnosed as "conversion" or "hysteria" has shown that the majority of studies reported misdiagnosis rates of <10%. Studies published from 1980 showed a mean rate of misdiagnoses of 4% (30). For somatization disorder, one of 37 patients with this disorder was detected as misdiagnosed 4 years later (13). In one of our studies, 6 (2%) of 295 primary care patients with multiple somatoform symptoms were found as possible misdiagnosed organic condition during follow-up (27). Although misdiagnoses rates should always be minimized, for somatoform disorders this rate is in the range of misdiagnoses that are found for other mental or physical disorders.
Implications for Revising the Somatoform Disorders Category to DSM-V
The existing criteria for somatoform disorders are an issue for debates (31–33), and most experts agree that there is a substantial need for improving the classification of people with multiple somatic complaints not better explained by a known medical condition. Whereas some authors favor an abolition of this category and moving these diagnoses to other mental or physical disorders, others argue to improve the existing category.
This overview on stability aspects of classification criteria offers some aspects that help to improve the quality of these diagnoses. The implications for the planning of future classification systems are outlined below.
Focusing on Multiple Somatic Complaints
Despite the critique of some experts (31,32) on the category of somatoform disorders in general, we suggest continuing to focus on somatic complaints when classifying these disorders, and not to merge this category under affective or anxiety disorders. The number of bodily symptoms is a relevant predictor of outcome in many different medical settings, and predicts certain outcomes (e.g., functional status, health care use) as well or better than does anxiety or depression (34,35). Moreover, somatoform symptoms have a specific impact on health services use that is not explained by depression or anxiety (36,37). Further, somatic complaints show different treatment responses in antidepressant trials compared with pure depression or anxiety disorders (38,39).
Results on stability of somatic complaints show that individual symptoms vary tremendously over time, whereas polysymptomatic syndromes have stable features. This is a reason to distinguish monosymptomatic and polysymptomatic syndromes in the classification. Different proposals for polysymptomatic somatoform disorders have been presented (11,35,40). All of these diagnostic groups show considerable impairment (28), thus justifying classification as a disorder. However, the cut-offs for classification should be much lower than previously used for somatization disorder, as the diagnosis of somatization disorder covers only a very small part (<5%) (41) of patients seeking medical help for multiple somatoform symptoms. However, as soon as more than two symptoms are present, the syndrome seems to be stable over time (23).
Although we favor a continuation of focusing on somatic complaints, we also feel that somatic complaints are not sufficient to justify a DSM-axis-1 or an International Classification of Disease (ICD)-10 section F diagnosis (42). Therefore, further psychological and behavioral features must be included to underline the relevance of these disorders under a psychosocial perspective. Examples for psychological and behavioral criteria can include symptom cognitions (selective attention, catastrophizing, focusing on organic explanations of the symptoms, expecting persistence of symptoms), affective aspects (e.g., demoralization, negative affectivity), or behavioral features (body checking, avoidance behavior, seeking for repeated testing). Most of these features have been shown to be specific for somatoform disorders (43), and thus, fulfill the necessity of classification criteria to be empirically based.
Present State Instead of Lifetime Diagnoses
Considering the weakness associated with remembering long time periods, the reliability and validity of these diagnoses can be improved by focusing on present state symptoms. However, this does not mean that time issues should be omitted completely. Persistence of symptoms is still a feature increasing the stability of the diagnosis, and knowing about similar former symptoms can help to increase diagnostic confidence. However, the strong dependency on time issues >2 years, as suggested in ICD-10 somatization disorder, or beginning of symptoms before the age of 30 years, as required for DSM-IV somatization disorder, are unnecessary sources of unreliability. The impact of these time-based criteria should be reduced.
Reduce the Impact of Doctors’ Etiology Assumptions Substantially
Doctors’ ratings that symptoms are "medically unexplained" is highly problematic. Physicians seem to differ substantially in the way they rate symptoms as due purely to organic processes versus psychologically influenced. Therefore, the impact of this source of variation should be reduced. In other examples, DSM-IV-TR uses the term "Symptoms are not better accounted for by another mental or physical disorder." This seems to be less dependent on doctors’ causality ratings, although this should be tested in more detail. Moreover, the potential positive role of knowing about the medical records (e.g., history of symptoms, investigations, and treatments) might further improve diagnostic validity. Diagnosis should not only represent the result of a negative selection of excluded medical conditions; instead, positive and specific criteria to define the disorder must be included (42). This again is a plea for further psychological and behavioral classification criteria.
Summary
Although there is a substantial need to improve existing classification approaches for somatoform disorders, the reported results on stability aspects of classification criteria indicate ways to improve it. The health care relevance of these disorders necessitates feasible diagnoses with high interrater reliability. A diagnosis for patients with multiple somatic complaints is needed and the inclusion criteria for this diagnosis should be broader than for somatization disorder. As these syndromes are under the section F (ICD-10) and axis 1 of DSM, they should not only depend on physical symptoms but also on psychological and behavioral features characterizing these patients.
The mentioned studies from Dr. Rief et al. were funded from the German Ministry of Research and Education (BMBF). There were no significant other sources of funding or support that might have influenced the content of this manuscript.
NOTES
Received for publication January 19, 2007; revision received June 25, 2007.
This article is being co-published by Psychosomatic Medicine and the American Psychiatric Association.
DOI:10.1097/PSY.0b013e31815b006e
REFERENCES
This article has been cited by other articles:
 |
|
 |
|
  R. C. Smith, J. C. Gardiner, Z. Luo, and K. Rost Diagnostic Accuracy of Predicting Somatization From Patients' ICD-9 Diagnoses Psychosom Med, April 1, 2009; 71(3): 366 - 371. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. olde Hartman, L. Hassink-Franke, C Dowrick, S Fortes, C Lam, H. van der Horst, P. Lucassen, and E. van Weel-Baumgarten Medically unexplained symptoms in family medicine: defining a research agenda. Proceedings from WONCA 2007 Fam. Pract., August 1, 2008; 25(4): 266 - 271. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Jackson and K. Kroenke Prevalence, Impact, and Prognosis of Multisomatoform Disorder in Primary Care: A 5-Year Follow-up Study Psychosom Med, May 1, 2008; 70(4): 430 - 434. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. E. Dimsdale, V. Patel, Y. Xin, and A. Kleinman Somatic Presentations A Challenge for DSM-V Psychosom Med, November 1, 2007; 69(9): 829 - 829. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
