Johnson, Susan K. PhD; DeLuca, John PhD; Natelson, Benjamin H. MD
According to DSM-III-R (), the diagnosis of somatization disorder (SD) requires a person to present with at least 13 symptoms for which no significant organic pathology can be found. The symptoms must have caused the person to take medication, to see a physician, or to have altered her/his lifestyle. The disorder begins before the age of 30 and has a chronic but fluctuating course. However, the diagnosis of SD is extremely problematic in terms of its validity because it involves a series of judgments that can be arbitrary and subjective (). Specifically, the interviewer must decide if the symptom reported is attributable to an identifiable medical illness. Although such judgments are extremely difficult to make uniformly, the influence of bias introduced by the interviewer's orientation on the prevalence of SD has not been adequately addressed.
The chronic fatigue syndrome (CFS) presents an excellent illustration of the problems inherent in the diagnosis of SD. The Centers for Disease Control (CDC) developed a clinical case definition of CFS () that requires new onset of 6 months of severe fatigue in a previously healthy person. In addition, the patient must report eight of the following problems: mild fever, sore throat, painful lymph nodes, muscle weakness, myalgia, severe fatigue after mild exercise, headaches, arthralgia, neuropsychological complaints, sleep disturbance, and acute onset of symptoms. The CDC criteria were developed to provide some uniformity for research studies, but the independent validity of these criteria has not been established. The concept of SD has a longer research history and greater construct validity than that of CFS. A number of family studies using the earlier construct of Briquet's syndrome has established reasonable reliability (), validity (), diagnostic stability (), and positive familial history (). More recently, SD has been shown to be associated with increased health service utilization () with an increased prevalence in women ().
Whether or not symptoms of CFS are considered medically caused will strongly affect the incidence of SD within the CFS population. Because arguments exist as to whether CFS is an organic disease or an emotional disorder ([10,11]), chances for inconsistencies in diagnosis are high. If the examiner recognizes that the patient's CFS symptoms indicate a physical illness, the diagnosis of SD may not be made. Conversely, if the examiner does not consider CFS a medical illness, the patient's symptom endorsement may lead to the diagnosis of SD.
Although prevalence of SD is below 0.03% in community samples (), prior studies have found relatively high, albeit variable, rates of SD in patients complaining of fatigue. In studies of patients who meet the CDC criteria for CFS, the rates of SD have ranged from 0 to 28% ([13-16]), and, in patients with fatigue who did not meet or were not assessed with CDC criteria, the range is from 0 to 20% ([16-20]) (Table 1). The large variation among these studies is in itself indicative of the problem in defining SD. Several factors that most likely contribute to this variability are: the different instruments used to assess SD, the interviewer's orientation, which influences how symptoms are coded (even when the same instrument is used), the heterogeneity in the fatigue populations studied, the lack of specificity on how fatigue-related symptoms were scored, and the lack of information on how "functional" diagnoses were scored.
The use of different diagnostic instruments contributes to variability in SD prevalence. For example, in prior studies using the Diagnostic Interview Schedule (DIS) to compare fatigue patients to those with similar medical illnesses ([18,19,21]), higher rates of SD were found in the fatigue groups than in studies using the Structured Clinical Interview for DSM-III-R (SCID) ([14,20]) (Table 1). This lower prevalence may be due to the fact that the SCID relies on clinical judgments that are relied on less when using highly structured instruments such as the DIS.
Another factor contributing to variability in prior studies assessing SD in CFS is the heterogeneity of the population. Prior studies have included patients with a high incidence of lifetime and current psychiatric diagnoses, most often depression or anxiety disorder. Such individuals may skew the results because of the increased risk for patients with depression and anxiety to somatize ([22-27]).
Lack of information as to how SD symptoms are coded is another factor contributing to variability. Only two prior studies ([21,26]) have specified which SD symptoms were considered to overlap with CDC symptom criteria for CFS. The difficulty in distinguishing among somatic symptoms that are psychiatric vs. organic in origin can result in overdiagnosis of SD in medical illness, particularly chronic illness. For example, Lustman et al. () performed a recalculation of the DIS diagnosis of SD by counting symptoms of diabetes mellitus as psychiatric symptoms and found that SD increased from 1.7 to 4%. Several studies ([13,15,16]) have performed a similar recalculation, changing CFS from a psychiatric to a physical disorder, which greatly reduced the number of subjects with SD (this is illustrated in Table 1). The present study examined how changing criteria for coding of SD symptoms affects rates of diagnosing SD.
The primary purpose of the present study was to examine how the number of medically unexplained (psychiatric) symptoms in CFS patients changes when different criteria sets are used to explain SD symptoms. The present study attempted to control for many of the factors that have contributed to variability in diagnosing SD in CFS. To reduce heterogeneity, only patients who met CDC criteria for CFS at the time of intake and had no psychiatric history in the 5 years preceding illness onset were included. To clarify how judgments were made, we have explicitly stated the coding format used with the DIS, i.e., what symptoms are coded as "psychiatric," including which "functional" disorders were coded as psychiatric. We then recalculated the number of psychiatric symptoms based on different criteria for judging psychiatric vs. medically explained symptoms. Additionally, the relative frequency of SD symptoms and the prevalence of SD was examined within two comparison groups of illnesses that have fatigue as a significant component, a "mild" multiple sclerosis (MS) group in which fatigue is thought to be neurologically based and a depressed group in which fatigue is thought to be psychiatrically based. A third control group was comprised of healthy but sedentary individuals.
Subjects consisted of 42 CFS, 33 healthy, 18 MS, and 21 depressed (DEP) individuals who were evaluated at the CFS Center at New Jersey Medical School's outpatient facility. The groups did not differ statistically in mean age and level of education (Table 2). Healthy and DEP subjects were recruited from advertising in the University community and surrounding area and were paid for their participation. Healthy and DEP subjects who reported no medical problems and were taking no medication other than birth control pills were included as subjects. DEP subjects had to be diagnosed with either major depression or dysthymia according to the DIS DSM-III-R criteria. MS subjects were referred by a neurologist and were classified with clinically definite MS (). Inclusion criteria consisted of a Kurtzke Expanded Disability Status Score (EDSS) () of 2.5 or below and no health problems other than MS. The mean EDSS score for the MS group was 1.2 (SEM = 0.18). Patients with such low EDSS scores are often asymptomatic and usually show no outward evidence of their illness.
Potential CFS subjects were accepted as research participants after a careful history, physical examination, and elimination of possible medical causes of fatigue ([3,31]). To be accepted as subjects, CFS patients had to report sustained fatigue severe enough to have reduced activity by at least 50%, and the fatigue had to last at least 6 months and be unexplained by other known medical causes. Also, patients had to endorse 8 of the 11 minor symptoms whose presence are required as part of the CDC case definition for CFS. To be counted, these symptoms had to have produced a significant problem in the patient's life in the month before their visit. To improve accuracy in recalling events relative to illness onset and to avoid the influence of psychosocial factors related to chronic illness, only those subjects who fulfilled the above criteria for CFS and had an illness duration of less than 4 years were included.
Subjects with a history of loss of consciousness for greater than 5 minutes, substance abuse, eating disorders, schizophrenia, or bipolar disorder were excluded from the study. CFS and MS subjects with affective or anxiety disorders concurrent with their illness or that had occurred more than 5 years before illness onset were not excluded. Two of the CFS subjects had a prior major depressive episode, one MS subject had posttraumatic stress disorder, one MS subject had a prior phobia, and one MS subject had a prior major depressive episode, but all of these conditions had resolved at least 5 years before illness onset.
Psychiatric history was obtained using a structured psychiatric interview, the computerized version of the Diagnostic Interview Schedule (Q-DIS) administered by a psychologist or neuropsychology technician trained in its use. To obtain a more detailed account of SD than that available on the Q-DIS, the paper and pencil version of the DIS III-R () somatization section was given. In the SD section, to qualify as possibly psychiatric, a physical symptom must cause a person to seek a physician or other health professional and to take medication more than once; the symptom must interfere with activities and have no organic (medically substantiated) explanation. Additionally, the symptom cannot be an effect of medication or drug or alcohol use. If these conditions are met, the symptom is coded as psychiatric. The DIS SD section contains 48 symptom questions, 38 of which are DSM-III-R symptom questions (the remaining 10 are from other diagnostic scales).
To examine the issue of variability in coding symptoms as psychiatric vs. physical in the diagnosis of SD, the DIS results were interpreted using four different criteria sets. Each set represents a different approach for judging SD symptoms as psychiatric or physical.
Set 1 consisted of recording the total number of 48 SD symptoms on the DIS, which were coded as psychiatric. SD symptoms were coded as psychiatric when the following explanations were given by a health professional: stress, anxiety, depression, "nerves," mental illness, or no definite diagnosis. If the subject did not contact a health professional about the symptom, but either took medication or changed normal activities because of it, it was coded as psychiatric.
Psychiatric coding was also used for the following medical explanations of a symptom. These are conditions that are often considered functional: CFS, Epstein-Barr virus, fibromyalgia, premenstrual syndrome, migraines, irritable bowel syndrome, fibrositis, hypoglycemia, vague food allergies, mitral valve prolapse, and candidiasis (yeast). The following were scored as psychiatric if there were no physical findings to corroborate the condition: dysmenorrhea, gastritis, and chronic back pain. From this point on, "SD symptom" will refer to any symptom that was coded as psychiatric according to these criteria.
Set 1 was broken into Sets 1a and 1b based on different cutoffs for diagnosis of SD. Set 1a used a cutoff of six SD symptoms for women and four for men (subsyndromal SD ()), and Set 1b used 13 SD symptoms (full SD) as a cutoff.
In this subset, the total number of 48 SD symptoms on the DIS that were coded as psychiatric was determined. This total was then used to determine the prevalence of SD under Escobar's subsyndromal somatization construct. Because full DSM-III SD is rare in the community, Escobar et al. () proposed using six SD symptoms for women and four for men as a cutoff for the diagnosis of SD. This abridged somatization construct was associated with similar risk factors, service use patterns, and disabilities as the full SD diagnosis ().
In this subset, the total number of 48 symptoms on the DIS that were coded as SD was recorded, and the prevalence of SD was determined using the criterion of 13 or more SD symptoms.
Set 2 viewed the CDC criteria CFS symptoms as "organic;" thus, CFS symptoms were not coded as psychiatric. In Set 2, headaches, arthralgia, myalgias, periods of muscle weakness, having to take time off due to unexplained illness, problems with walking, crying spells, and being bothered for 6 months or more by pains were counted as the CDC criteria CFS symptoms. (It should be noted, however, that headaches, time off due to unexplained illness, and 6 months or more of pains are not part of the DSM-III-R list of SD symptoms.)
Set 2 was also broken down into 2a (subsyndromal SD) and 2b (full SD). Set 2a used a cutoff of six SD symptoms for women and four for men, and Set 2b used 13 SD symptoms as a cutoff.
In Set 3, in addition to the Set 2 symptoms coded as physical, all symptoms that have occurred since the onset of CFS and that the patient considers due to CFS were coded as physical. Thus, psychiatric symptoms in Set 3 were only those SD symptoms that the subject experienced before CFS onset. A similar manipulation was used for the other two illness groups, i.e., for the MS group only SD symptoms that occurred before MS diagnosis and for the DEP group only SD symptoms that occurred before the current episode of depression counted toward the full SD diagnosis.
Set 4 used strict DSM-III-R criteria to determine prevalence of SD. Thus, only individuals with onset of CFS before age 30 who had 13 or more SD symptoms were considered to had SD. Set 4 used only the 39 DSM-III-R criteria symptoms, and CDC criteria CFS symptoms were coded as physical (as in Set 2).
The prevalence of SD across the various set conditions are presented in Figure 1. When the total number of SD symptoms are used (Sets 1a and 1b), significantly more CFS subjects meet the diagnosis of SD than subjects in the other three groups. Specifically, on Set 1a (subsyndromal SD),2 analysis revealed significant differences in frequency of SD cases between the CFS and Healthy (2) = 66, p <.0001), CFS vs. MS (2) = 38, p <.0001), and CFS vs. DEP (2) = 22.7, p <.0001). There were also significant differences between the CFS and Healthy (2) = 27, p <.0001), CFS and MS (2) = 16.4, p <.0001), and CFS and DEP (2) =13.1, p <.001) groups on Set 1b (full SD).
When CDC criteria CFS symptoms are excluded from the diagnosis of SD (Sets 2a and 2b), CFS was significantly different from Healthy (2) = 39, p <.0001) and MS (2) = 13.8, p <.001) but not from DEP (2) = 1.7, p <.18) on Set 2a (subsyndromal SD). On Set 2b (full SD), CFS was significantly different from Healthy (2) = 4.1, p <.05) but not from MS (2) = 2.4, p <.13) or DEP (2) =.08, p <.78).
There were no significant differences between the Healthy and MS nor the Healthy and DEP groups on full SD. However, using the subsyndromal SD criterion, there was a higher frequency of SD cases in the MS vs. Healthy groups (2) = 4.7, p <.03) and the DEP group vs. Healthy group (2) = 15.3, p <.0001).
No subjects in any of the groups received a full SD diagnosis in Set 3. This set excluded SD symptoms that occurred after onset of illness in the CFS, MS, and DEP groups, suggesting that for these groups SD symptoms are associated with illness onset and are not necessarily a part of a lifetime pattern of somatization.
On Set 4, using the strict DSM-III-R definition of SD, only one CFS and one DEP subject were identified with SD (Figure 1). DSM-IV has made the diagnosis of SD more specific by reducing the number of symptoms while requiring that symptoms cover a breadth of organ systems. However, reanalysis of the data using DSM-IV criteria did not substantially change the results. *
(Table 3) presents the percentage of symptoms on the DIS that were endorsed by at least 10% of any group in this study. With respect to CDC criteria symptoms, these items were endorsed primarily by CFS subjects, with endorsements ranging from 40% (crying spells) to 95% (muscle weakness) of the CFS sample. Crying spells was the only CDC minor symptom in which endorsement was greater in a group (ie, DEP) other than CFS. It is notable that the CFS group endorsed the highest percentage of symptoms among the four groups in 17 of 26 non-CDC symptoms in Table 3. Thus, 76% of CFS subjects endorsed "dizziness," 55% endorsed "shortness of breath" and "tachycardia," and 43% endorsed "nausea" and "loss of feeling." These percentages were as high as many CDC criteria symptoms for CFS.
An effort was made to determine which of the non-CFS groups most resembled the CFS group in symptom endorsement. Non-CDC symptoms in Table 3 that were endorsed by at least 30% of the CFS group were examined. Of the 11 symptoms identified, the DEP group appeared most like the CFS group in 9 of the 11 symptoms. Thus, both the CFS and DEP groups tended to endorse a greater percentage of DIS symptoms than either the MS or Healthy groups.
(Table 4) presents concurrent DIS diagnoses other than SD for the three illness groups. A total of 25 (60%) of CFS subjects had no other DIS diagnosis, and 16 (40%) had other DIS diagnoses that began at or after CFS onset. An analysis was conducted to examine whether CFS subjects with DIS diagnoses differed in the mean number of SD symptoms endorsed from CFS subjects without a DIS diagnosis. No statistically significant differences were observed between the two groups for Set 1 (F(1,38) = 2.4, p =.13), Set 2 (F(1, 38) = 1.6, p =.21), or Set 3 (F(1,38) =.14, p =.71) SD symptoms.
The homogeneity of a patient population is critical when trying to evaluate a controversial syndrome such as CFS. The present study used the most homogeneous group of CFS patients of any study to date. Even with this reduced variability, prevalence rates of SD ranged from 0 to 98% depending on whether CFS symptoms were coded as being due to a physical illness or not (Figure 1). Thus, the judgments made regarding whether a symptom is coded as physical or psychiatric dramatically affect prevalence rates of SD in CFS.
Three prior studies have also reported a reduction in SD prevalence when CDC criteria CFS symptoms were coded as physical. The present study found a reduction from 55 to 12% in SD rate when CDC criteria CFS symptoms were coded as physical (compare Sets 1b and 2b). The rate of SD prevalence in the subjects of Kreusi et al. () dropped from 20 to 10%, those of Demitrack et al. () from 10 to 3%, and those of Lane et al. () from 28 to 12% when CFS symptoms were coded as physical. These relatively high rates of SD, even after removing CDC criteria CFS symptoms from the equation, indicate that CFS subjects report many non-CFS symptoms that are then coded as psychiatric by the individual collecting the information.
Using the strict DSM-III-R definition identifies very few CFS patients as having SD (Set 4). This is in contrast to using either the subsyndromal definition of SD or a symptom count of greater than or equal to13 SD symptoms. Thus, although CFS patients frequently endorse symptoms thought to indicate SD, they do so only after the onset of CFS and thus do not show the history of sickliness thought to be characteristic of SD. If CFS is considered an organic disease, then CFS patients cannot have SD. The finding that CFS patients report many other symptoms besides those developed in the CDC case criteria might then suggest that those criteria need to be amended to include more symptoms.
Alternatively, results of the present study indicate that, although few individuals with CFS meet formal DSM-III-R criteria for SD, this does not preclude them from "somatizing." A large body of literature demonstrates that individuals with depression and anxiety ([22-27,33]), negative affectivity (), and high neuroticism () report large numbers of somatic symptoms. Furthermore, primary care utilization is doubled in patients with mental disorder vs. those without (). Kroenke and Price () found that increased prevalence of somatic symptoms in the general population was associated with an increased lifetime prevalence of psychiatric disorder. Katon and Russo () also found that chronic fatigue patients with the greatest numbers of symptoms had the highest rates of current and lifetime psychiatric disorder. In the present study, CFS subjects reported a very high number of somatic symptoms across a variety of organ systems, and these symptoms paralleled many of those reported by the depressed group. The DEP group also showed a tendency toward somatization, although only one subject was diagnosed with SD using Set 4. Table 3 illustrates the similarity in the type of symptoms reported and the percentage of those symptoms between the CFS and DEP groups. The CFS and DEP groups endorsed substantially more of the non-CDC symptoms than either the MS or Healthy groups, including: abdominal pain, chest pain, nausea, excess gas, blurred vision, shortness of breath, dizziness, and a lump in the throat. These results indicate that, in addition to CDC criteria symptoms, CFS is characterized by a general increase in somatic symptoms not necessarily reflective of the infectious/immunological disease process that the CDC criteria target.
Contrary to the existing literature and the parallel symptomatology with the depressed group, however, is the present finding that the CFS subjects with depression and anxiety disorders did not report significantly more SD symptoms than CFS subjects with no DIS diagnoses. Furthermore, screening out chronic psychiatric disorders in our population did not result in a decrease in somatic symptom reporting compared with previous CFS studies. These results strongly suggest that psychiatric factors alone are not a sufficient explanation for the broad array of somatic symptoms reported in CFS.
CFS can be viewed as an organic disease involving many organ systems or an undifferentiated somatoform disorder. Arguing against the former is the lack of consistent pathophysiological findings in these patients; arguing against the latter is the sudden onset of their symptoms with no history of somatizing behavior. However, a less dichotomous and more likely possibility is that the CFS population is heterogeneous, consisting of individuals in both categories. Future research in CFS needs to focus on defining homogenous subgroups of CFS to determine whether contributions to the disease process are more heavily organic or psychological in different subgroups. For instance, in the present study, we excluded individuals with chronic psychiatric disorder; including these individuals and comparing them with a nonpsychiatric CFS subgroup could mitigate some of the ambiguity surrounding the CFS literature. This should help refine the search for etiological agents and identify the effectiveness of differential treatment approaches.
The present study illustrates that the terminology used to interpret the symptoms (ie, psychiatric or physical) will determine which category CFS falls into, yet neither language advances the field. A diagnosis of SD may be so arbitrary as to be rendered meaningless in controversial illnesses such as CFS. This problem has been addressed in Graham's work on linguistic parallelism () positing that "functional" vs. "organic" are useless concepts reflecting different terminology used to explain the same event. Graham emphasized that psychological mechanisms do not exclude physical ones and that many symptoms dismissed as "functional" have pathophysiological mechanisms. This physiological basis for common somatization symptoms has been supported recently by the work of Sharpe and Bass (). Similarly, Kirmayer and Robbins () hypothesize that the distinction between functional and organic is that between process and structure, with functional disorders involving abnormal processes in structurally intact organ systems.
Labeling of a condition is not trivial because it can affect treatment services. Our present state of knowledge indicates that treatment for patients with controversial illness involving many medically unexplained symptoms should be conservative because "overinvestigation" consumes health care resources without benefiting the patient ().
This research was supported in part by Grant U01AI-32247 from the National Institutes of Health, Bethesda, MD, establishing a Chronic Fatigue Syndrome Research Center at the University of Medicine and Dentistry of New Jersey-New Jersey Medical School, Newark, NJ.
Dr. Johnson was supported by Grant H133 P10002-92 from the Department of Education, National Institute of Disability and Rehabilitation Research, Bethesda, MD, to the University of Medicine and Dentistry of New Jersey-New Jersey Medical School.
The authors thank Dr. Stuart Cook for referring the MS patients studied here. The authors also thank Nancy Hill for her administrative work, Michael Kresloff for data analysis, and Dawn Beldowicz, Steve Kovel, and Jackie Pareja for their help in administering the DIS interview.
* Reanalysis by DSM-IV resulted in 31% of CFS group diagnosed with SD by Set 1, 8.4% by Set 2, 0 by Set 3, and 2.3% by Set 4 criteria.
1. Diagnostic and Statistical Manual of Mental Disorders,. 3rd Ed-Revised Washington, DC, American Psychiatric Association, 1987
2. Kellner R: Psychosomatic syndromes, somatization and somatoform disorders. Psychother Psychosom 61:4-24, 1994
3. Holmes GP, Kaplan JE, Gantz NM, et al: Chronic fatigue syndrome: A working case definition. Ann Intern Med 108:387-89, 1988
4. Spitzer R, Endicott J, Robins E: Research diagnostic criteria: Rationale and reliability. Arch Gen Psychiatry 35:773-782, 1978
5. Guze S: The validity and significance of hysteria (Briquet's Syndrome). Am J Psych 132:138-141, 1975
6. Perle MG, Guze SB: Hysteria: The stability and usefulness of clinical criteria. N Engl J Med 266:421-426, 1962
7. Cloninger C, Reich T, Guze S: The multifactorial model of disease transmission. III. Familial relationship between sociopathy and hysteria (Briquet's Syndrome). Br J Psychol 127:11-22, 1975
8. Escobar JI, Golding JM, Hough RL et al: Somatization in the community: Relationship to disability and use of services. Am J Public Health 77:837-840, 1987
9. Cloninger C, Martin R, Guze S, et al: Somatization disorder in men and women: A prospective follow-up and family study. Am J Psychol 143:873-878, 1986
10. Manu P, Lane TJ, Matthews DA: Chronic fatigue and chronic fatigue syndrome: Clinical epidemiology and etiological classification. In Chronic Fatigue Syndrome, Chichester, England, Wiley, 1988
11. Stewart DE: Emotional disorders misdiagnosed as physical illness: Environmental hypersensitivity, candidiasis hypersensitivity, and chronic fatigue syndrome. Int J Ment Health 19:56-68, 1990
12. Escobar JL, Burnam A, Karno M, et al: Somatization in the community. Arch Gen Psychiatry 44:713-18, 1987
13. Kruesi MJP, Dale J, Straus SE: Psychiatric diagnoses in patients who have chronic fatigue syndrome. J Clin Psychiatry 50:53-6, 1989
14. Hickie I, Lloyd A, Wakefield D, et al: The psychiatric status of patients with the chronic fatigue syndrome. Br J Psychiatry 156:534-40, 1990
15. Demitrack MA, Dale JK, Straus SE, et al: Evidence for impaired activation of the hypothalamic pituitary-adrenal axis in patients with chronic fatigue syndrome. J Clin Endocrinol Metab 73:1224-34, 1991
16. Lane TJ, Manu P, Matthews DA: Depression and somatization in the chronic fatigue syndrome. Am J Med 91:335-44, 1991
17. Manu P, Lane TJ, Matthews DA: Somatization disorder in patients with chronic fatigue. Psychosomatics 30:388-95, 1989
18. Wessely S, Powell R: Fatigue syndromes: A comparison of chronic "postviral" fatigue with neuromuscular and affective disorder. J Neurol Neurosurg Psychiatry 52:940-48, 1989
19. Wood GC, Bentall RP, Gopfert M, et al: A comparative psychiatric assessment of patients with chronic fatigue syndrome and muscle disease. Psychol Med 21:619-28, 1991
20. Pepper CM, Krupp LB, Freidberg F, et al: A comparison of neuropsychiatric characteristics in chronic fatigue syndrome, multiple sclerosis, and major depression. J Neuropsychiatry Clin Neurosci 5:200-05, 1993
21. Katon, WJ, Buchwald DS, Simon GE, et al: Psychiatric illness in patients with chronic fatigue and those with rheumatoid arthritis. J Gen Intern Med 6:277-85, 1991
22. Escobar JI, Canino G: Unexplained physical complaints: Psychopathology and epidemiological correlates. Br J Psychiatry 154:24-7, 1989
23. Kirmayer LJ, Robbins JM, Drorkird M, et al: Somatization and the recognition of depression and anxiety in primary care. Am J Psychiatry 150:734-41, 1993
24. Katon W: The epidemiology of depression in medical care. Int J Psychiatry Med 17:93-113, 1987
25. Katon WJ, Russo J: Somatic symptoms and depression. J Fam Pract 29:65-9, 1989
26. Katon WJ, Russo J: Chronic fatigue syndrome criteria: A critique of the requirement for multiple physical complaints. Arch Intern Med 152:1604-09, 1992
27. Lipowski ZJ: Somatization and depression. Psychosomatics 31:13-21, 1990
28. Lustman PJ, Harper GW, Griffith LS, et al: Use of the Diagnostic Interview Schedule in patients with diabetes mellitus. J Nerv Ment Dis 174:743-46, 1986
29. Poser CM, Paty DW, Scheinberg, et al: New diagnostic criteria for multiple sclerosis: Guidelines for research protocols. Ann Neurol 13:227-31, 1983
30. Kurtzke JF: Rating neurological impairment in multiple sclerosis: An expanded disability status scale (EDSS). Neurology 33:295-301, 1983
31. Schluederberg A, Straus SE, Peterson P, et al: Chronic fatigue syndrome research: Definition and medical outcome assessment. Ann Intern Med 117:325-31, 1992
32. Robins L, Cottler L, Keating S: Diagnostic Interview Schedule (DIS): Version III-Revised, St. Louis, MO, Washington University School of Medicine, 1991
33. Abbey SE, Garfinkel PE: Chronic fatigue syndrome and depression: Cause, effect, or covariate. Rev Infect Dis 13(Suppl 1):73-83, 1991
34. Krupp LB, Mendelson WB, Friedman R: An overview of chronic fatigue syndrome. J Clin Psychiatry 52:403, 1991
35. Katon W, Kleinman A, Rosen G: Depression and somatization: A review. Am J Med 72:127-135, 1982.
36. Watson D, Pennebaker JW: Health complaints, stress, and distress: Exploring the central role of negative affectivity. Psychol Rev 96:234-254
37. Costa PT, McCrae RR: Neuroticism, somatic complaints, and disease: Is the bark worse than the bite? J Pers 55:299-316, 1987
38. Hankin J, Oktay JS: Mental disorder and primary care: An analytic review of the literature. In National Institute of Mental Health, series D, No. 7 (DHEW Publication No. ADM 78:661). Washington, DC, US Government Printing Office, 1979
39. Kroenke K, Price RK: Symptoms in the community: Prevalence, classification, and psychiatric comorbidity. Arch Int Med 153:2474-2480, 1993
40. Graham DT: Health, disease, and the mind-body problem: Linguistic parallelism. Psychosom Med 24:52-71, 1967
41. Sharpe M, Bass C: Pathophysiological mechanisms in somatization. Int Rev Psychol 4:81-97, 1992
42. Kirmayer LJ, Robbins JM: Functional somatic syndromes. In Kirmayer LJ, Robbins JM (eds), Current Concepts of Somatization: Research and Clinical Perspectives. Washington, DC, American Psychiatric Press, 1991
43. Mayou R: Medically unexplained symptoms (editorial). Br Med J 303:534-535, 1991
somatization disorder, chronic fatigue syndrome, depression, fatigue.
Copyright © 1996 by American Psychosomatic Society