Journal of Behavioral Medicine, VoL 13, No. 6, 1990

Losing Face: Sources of Stigma as Perceived by Chronic Facial Pain Patients Joseph J. Marbach, 1.4 Mary Clare Lennon, I Bruce G. Link, 2 and Bruce P. Dohrenwend 3

Acceptedfor publication: March 13, 1990

The purpose o f this paper is to identify potential sources o f estrangement and feeling psychologically flawed as perceived by temporomandibular pain and dysfunction syndrome (TMPDS) patients. It is our hypothesis that a primary source of patients'perceived stigma results from pejorative labeling by clinicians. The data come from a study of 151 women TMPDS patients. The results show that the lack of a known etiology or pathogenesis for the condition allows the possibi6ty o f pejorative labeling by influential others such as physicians and dentists that in turn causes TMPDS sufferers to feel stigmatized. Stigmatization is not the result of clinical factors p e r se or personality problems. Perceived stigma associated with TMPDS leads to the same sorts of strained interactions and feeling estranged that have been shown to be typical of people with other stigmatized conditions. These strains and feelings in turn may contribute to ill health. KEY WORDS: temporomandibular pain dysfunction; facial pain; social stigma.

~Division of Sociomedical Sciences, School of Public Health, Columbia University, New York, New York 10032. 2Division of Epidemiology, School of Public Health, Columbia University, New York, New York 10032. 3Department of Psychiatry, Columbia University and New York State Psychiatric Institute, New York, New York 10032. 4To whom correspondence should be addressed at Division of Sociomedical Sciences, School of Public Health, Columbia University, 600 West 168th Street, New York, New York 10032. 583 0160-7715/90/1200-0583506.00/0 9 1990 Plenum Publishing Corporation

584

Marbach, Lennon, Link, and Dohrenwend

INTRODUCTION Chronic pain patients experience difficulties in social settings that result from unsolved problems in their clinical management (Hilbert, 1984; Kotarba, 1983). Lennon and colleagues (1989a) have located an important aspect of these problems in their analysis of the associations between perceptions of social stigma and strain in the intimate relationships of TMPDS patients. The present paper describes a further step in the form of an investigation of processes which may lead TMPDS sufferers to experience stigma as they seek care and treatment for their pain. By way of introduction to the empirical analyses, a brief review of the theory and clinical experience on which they are based is presented. Stigmatization is a complex process of social interaction between an unmarked or "normal" person and the bearer of an evident or presumed "mark" that defines the person as flawed, incomplete, spoiled, or undesirable. The mark may be physical such as being disfigured or may be attributed to behavior, biography, ancestry, or group affiliation (Jones et al., 1984). Those who are marked have in common that they often engender negative reactions in others that result in failure to accord the marked individual proper respect and regard in specific situations (Goffman, 1963). Part of the process of stigmatization involves stereotyping, that is, deriving generalizations about personal characteristics from inadequate and often superficial information. Once a flaw is suspected or identified, there is a tendency to impute a wider range of imperfections on the basis of the original one (Goffman, 1963). Health professionals are among those who stereotype, reject, or stigmatize and reportedly use the same criteria as others in the general society of which they are a part (Ford, 1962; Illich, 1976; Volinn, 1983). Such behavior on the part of professionals could be expected to have particular implications for TMPDS sufferers. TMPDS is a disorder of pain and dysfunction of the masticatory and associated muscles akin to other fibromyalgias (Merskey, 1986; Lipton and Marbach, 1983). When compared to nonpatient controls, TMPDS patients have elevated levels of distress, are beleaguered by physical illness and injuries, tend to attribute their fate to external factors, and have fewer sources of emotional support (Marbach et al., 1988). The disorder, often chronic, has no visible signs that distinguish cases from noncases. Nevertheless, carefully specified diagnostic criteria for TMPDS, though not etiologically based, are available (Merskey, 1986). Those who seek care for TMPDS are, for unknown reasons, predominantly adult women (Greene and Marbach, 1982). Prevalence rates of the syndrome are unknown. Moreover, high rates of the signs and symptoms of the disorder are reported in nonpatient populations by some (Kleinknecht et al., 1987) but not by others (Nuprin, 1985). Most of our

Sources of Stigma for Facial Pain Patients

585

knowledge of T M P D S is derived from observation of those who seek care. Little is known about which factors influence certain sufferers not to seek care or about whether those who report some of the signs and symptoms of TMPDS in epidemiologic studies would be considered cases using clinical criteria. Regardless of whether or not the reports of high rates of signs and symptoms in the general population are correct, it would appear that the syndrome is validated as a "disorder" not only by signs and symptoms, but by the act of care-seeking (Marbach and Lipton, 1978; Mechanic, 1980; Speculand etal., 1984) and obtaining a diagnosis (labeling) by an expert (Link, 1982, 1987). This validation process applies to all morbid conditions but is especially important to patients when the other criteria legitimizing their illness consist mainly of symptom reports and their own care-seeking behavior. If TMPDS sufferers experience stigma as previous research indicates (Lennon et al., 1989), the question arises as to where the process of stigmatization begins and how the stigma is sustained. TMPDS does not have a widely recognizable cultural image. Publicly labeling someone a "TMPDS patient" does not arouse the same kinds of sentiments as does, for example, labeling someone an "alcoholic," "child abuser," or "AIDS patient." In addition, there is no outward blemish that can directly elicit fear or disgust as might occur with psoriasis, or facial disfigurement. In fact, pain per se is not stigmatizing. For example, cancer patients or those suffering dental abscesses are unlikely to experience stigma as a result of their pain. One source of stigma associated with TMPDS may be the uncertainty of its cause; a second, the high rates of treatment failure (Lipton and Marbach, 1984a). These two factors are highly correlated in clinical practice and have implications for stigma. For example, because tooth grinding is considered by some to be a principal cause of TMPDS (Laskin, 1969), many patients are told by clinicians that they have TMPDS because they grind their teeth (Marbach et al., 1988, 1990a). However, treatment strategies such as oral appliances, designed to control the effects of tooth grinding, have proven frequently ineffective especially for the more severe cases of TMPDS (Clark, 1984; Feinmann, 1983). Moreover, the belief that tooth grinding causes T M P D S increases the potential for stigma because many clinicians believe that the underlying cause of tooth grinding is a psychological abnormality (Rugh and Harlan, 1988). Consequently, TMPDS is considered by many clinicians to be primarily a psychologically created disorder (Marbach et al., 1978; Salter et al., 1983). Other potential sources of stigma are related to what is perceived as an excessive use of health services by T M P D S patients (Von K o r f f et al., 1988) and their greater reliance on sedatives and analgesic drugs compared to other chronic pain patients (Swanson et al., 1986). Chronic pain patients are aware that use of these drugs is often regarded as the result of a personal

586

Marbach, Lennon, Link, and Dohrenwend

flaw or weakness and do not always provide treating clinicians with an accurate history of drug use. The pattern of underreporting is more pronounced for narcotics than tranquilizers (Ready et al., 1982). This underreporting could be a reaction to perceived stigma and is not surprising given the strong negative image associated with narcotic use. Indeed, patients are not the only ones who risk stigmatization when identified with advocating narcotic analgesic treatment. Clinicians who have suggested that there is a role for long-term use of narcotic analgesics for the treatment of certain chronic pain patients have been criticized for this practice on the grounds that it can prolong pain behavior (Sternbach, 1987). Government regulations add to the potential stigma associated with narcotic use. At the time this study was conducted, New York State officials required special, readily identifiable prescription forms solely for narcotic analgesics. Following the adoption of this procedure, obtaining narcotic analgesics from clinicians became more difficult. Against this background, one plausible scenario for events and conditions leading to stigmatization might be the following. Due to TMPDS's unexplained cause and chronicity, and motivated by pain, those afflicted conduct extensive searches for possible determinants. This search process itself can cause a profound questioning of the entire repertory of habits that give structure and meaning to a person's life. The individual becomes highly suggestible and hence vulnerable to the labeling inclinations of others. The opinions of physicians and dentists, as professionals who are officially sanctioned to evaluate and determine the source of facial pain, carry great weight. Let us spell out this last point in detail since it is pivotal for the present investigation. Evidence for the powerful associations between statuses in stratified social systems and well-being has been consistently demonstrated (Hollingshead and Redlich, 1958; Thoits, 1986). Sociological theory emphasizes that statuses and roles are found within the structural context of social institutions and their arrangements (Pearlin, 1989). For example, an employee has an employer, a wife a husband, and a patient a doctor. One does not act alone in a role. There is abundant evidence that these roles are relatively enduring, stable, and of considerable importance to the involved parties (Pearlin, 1989). Thus, when these relationships are strained, they can result in considerable distress. Chronic facial pain patients describe consultation with clinicians often as strained interactions (Marbach and Lipton, 1987). Unfortunately this is because when TMPDS patients look to these powerful persons for information, they are often provided with pejorative explanations (Salter et al., 1983). Patients discover that clinicians regard TMPDS as psychogenic (Feinmann, 1983) or as a depressive equivalent (Ford, 1983). Even if explicit psychological labels are avoided, patients may reach the same conclusion when they are required to plead for analgesics and are offered tranquilizers instead. Such

Sources of Stigma for Facial Pain Patients

587

negative feedback about themselves may lead the patient to feel marked or stigmatized, as others have noted (Hilbert, 1984; Kotarba, 1983). Chronic pain patients' complaints do not always conform to traditional notions of "normal pain" (Hilbert, 1984), which is seen as having a clearly biomedical cause and is usually of short duration (Merskey, 1988; Wall, 1984). By implication, identifying a particular pain disorder as psychogenic suggests that the sufferer may be responsible for the cause and/or maintenance of his/her pain (Angell, 1985; Cassell, 1982; Kotarba, 1983).

Hypotheses

If they are exposed to pejorative labeling by clinicians, TMPDS patients may experience stigma and be influenced in at least two ways. First and most obvious, the patients may develop a belief that others attribute TMPDS to flaws in their personalities. Second, TMPDS patients may come to feel a profound sense of being different, misunderstood, and alone (Lennon et al., 1989) when those who have the culturally assigned role of identifying the sources of their pain fail to do so to the sufferers' satisfaction. We hypothesize that such perceptions of being stigmatized will vary with the number of different clinicians a patient has been exposed to as well as with what the patient has been told about the cause of TMPDS by the clinician. The rationale for these hypotheses is that the more clinicians contacted, the greater the opportunity for stigmatizing labels to be applied and reinforced as the clinicians suggest that the source of the patient's pain is psychological, a "disorder of the mind" (Salter et al., 1983) or "imaginary." Furthermore, we believe that drugs such as narcotic analgesics and sedatives/tranquilizers serve as labeling variables in addition to their pharmacologic function. Namely, we hypothesize that because of the stigma associated with narcotic analgesics, particularly in relation to a chronic pain disorder, use of these drugs will be associated with manifestations of stigma. Two alternative hypotheses are investigated. First, stigma perceptions by the patient may result from cinical factors rather than from particular stereotyping reactions by clinicians. For example, the number of clinicians consulted is likely to vary directly with the duration and severity of pain. Because the pain often remains intractable, some patients may feel misunderstood and/or even come to blame themselves for a condition that does not yield to the "best efforts" of those who treat them. Second, the stigma perceptions of TMPDS patients may be associated with an underlying somatoform pain disorder (American Psychiatric Association, 1987) implying underlying personality problems. Thus, clinicians' reactions to TMPDS patients may accurately reflect a psychological basis for

588

Marbach, Lennon, Link, and Dohrenwend

the pain. If this hypothesis is correct, the patients' perceptions of stigma are the result of their psychological problems, rather than the stigmatizing behavior of the clinicians. The psychological problems would consist of persistent psychological distress, depression and anxiety, hypochondriasis, abnormal illness behavior, and so forth. This study tests empirically these alternative predictions and thereby assesses the plausibility of our thesis concerning the source of stigma in the labels applied to T M P D S patients by the clinicians they consult.

METHOD The Sample

The sample for this investigation comes from a study of life stress and pain. The cases were drawn from the list of a private practitioner who specializes in the treatment of chronic facial pain. To be eligible, the patients had to be white, English-speaking women and residents of the New York Metropolitan region. They also had to meet strict diagnostic criteria for TMPDS that are described shortly. Clinicians identified 201 patients who were eligible for the study on relevant clinical, residential, and demographic grounds. Of the 201 eligible patients, 199 were contacted and 151 (75o7o) of these entered the study. When those who participated in this interview were compared with those who refused, the two groups were found to be similar in average level o f education, age, health status, marital status, and suddenness of onset of TMPDS. The mean age of the patients was 38 years; the mean number of years of education completed was 15.2. Fifty-six percent of the patients were married or living with a mate, and 77O7o were employed for wages. The median family income was $44,285 (in 1983 dollars). Diagnoses of TMPDS were made according to the following widely accepted criteria (Merskey, 1986): (1) tenderness in one or more muscles of mastication, together with (2) clicking or popping noise in the temporomandibular joint (TMJ) a n d / o r (3) limitation of mandibular range. A lack of tenderness in the TMJ upon palpation by way of the external auditory meatus was not required. Exclusion criteria included (1) clinical and radiographic evidence of organic changes in the TMJ or positive neurological signs that could be associated with any other facial pain syndrome, (2) systemic inflammatory arthritis, or (3) infections or degenerative joint disease of the TMJ or ear infections. Also excluded were patients who were currently receiving care for a chronic disease other than TMPDS and those who had ever been inpatients in a psychiatric facility. The latter criterion could have resulted

Sources of Stigma for Facial Pain Patients

$89

in lower rates of major mental illness in this sample. No patients had undergone surgery of the TMJ. To test the reliability of the clinical TMPDS diagnostic schedule, three other clinicians independently reviewed records, including radiographs of the temporomandibular joints and laboratory tests, and applied these diagnostic criteria to a random sample of the study population (N = 6). There was perfect agreement among the four clinicians as to the diagnosis of TMPDS. Adequacy of Diagnostic Criteria

Although the criteria used for the diagnosis of TMPDS in this study are widely accepted (Merskey, 1986), the diagnosis of this and other myofascial pain syndromes is not without controversy about both physical (Osterweis et al., 1987) and psychological (Ahlers et al., 1984) indications. Specifically, with regard to TMPDS patients, attempts have been made to identify subgroups of TMPDS patients based on variability in TMJ anatomy, the "internal derangements" approach. This approach currently has a substantial following among clinicians but has less support in research data. Subgroups of TMD patients have not been found to differ on psychopathology, pain duration, and treatment outcome using structural abnormalities of the TMJ as criteria (Hellsing and Holmlund, 1985; Rudy et al., 1989). Only one study found small but significant differences on several MMPI subscales between different groups of facial pain patients (Eversole et al., 1985). However, no corrections were made for multiple comparisons, which suggests that the significant findings reported may not be very robust. Attempts to replicate these findings were unsuccessful (Rudy et al., 1989; Schnurr et al., 1990).

Representativeness of Cases

Data from the kind of epidemiological studies that are the prerequisites to securing ideal samples are not yet available. The relatively low prevalence of TMPDS in the general population makes identification of cases difficult. Evidence for the representativeness of the cases in this study, compared to TMPDS patients in general, comes from studies of patients seeking care in specialty clinics. Most of these patients were white, female, and middleand upper middle-class (Greene and Marbach, 1982), as are the cases in this study. Furthermore, the women in the present study differ little from TMPDS cases of different ethnicity, race, and class when compared on age, clinical, and psychosocial criteria as reported in previous studies, even though the latter were conducted at different times and in different treatment locations (Lipton and Marbach, 1984b; Kleinknecht et al., 1987; Heloe and Heloe,

590

Marbach, Lennon, Link, and Dohrenwend

1975). In short, the cases in the present study resemble patients included in previous studies even though the latter were conducted at different times and in different treatment locations.

The Measures Pain Measures. At the entrance interview, patients were asked to describe their pain at its worst in the previous month. Several types of ratings were used (Wallenstein et al., 1980), four of which form the basis of a composite pain severity score constructed for these analyses. The four measures are a direct magnitude estimation of overall pain severity (Sternbach et al., 1974), a visual analogue rating (Huskisson, 1983), the Pain Rating Index (PRI) derived from the McGill Pain Questionnaire (Melzack, 1975), and a categorical measure developed by Tursky (1974). Because these measures are highly intercorrelated, they were combined into a composite pain severity score, first converting the scoring of the PRI and categoric measures to have the 0-100 range of the other two measures. The composite pain severity score was then calculated by taking the average score on all four measures. This is a highly reliable indicator of pain severity (alpha = .93). As indicated in Table I, the average worst pain of the month severity score in this sample was 63.9 (SD, 23.3). We also assessed the duration of the pain by asking respondents to report how long ago their symptoms started. On average, patients in this study have had TMPDS for 6.4 years. The average age at onset of T M P D S was 32.0 years (Table I).

Labeling Variables Three variables indicative of labeling were developed. The first, the number o f clinicians consulted, was obtained by asking "After your symptoms began, who was the first person you consulted?" "Who did you consult next?" and so forth. As Table I shows, the average number of different clinicians consulted was 6.5 (SD, 3.7). The type of clinician (e.g., internist, dentist, otolaryngologist, chiropractor, faith healer, neurologist) and number of different consultations were recorded. Information on the type of clinician consulted is of interest because not all health providers are licensed to prescribe drugs. This could influence our findings o f drugs as labeling variables. On the first pass, cases reported consulting 148 different clinicians prior to consulting us. That is, we were the initial consultants for three patients. All but five clinicians, four chiropractors and one psychologist, were potential drug prescribers. Infor-

Sources of Stigma for Facial Pain Patients

591

Table I. Clinical Characteristics (N = 151)

Worst pain (range, 0-100) Duration (years) of pain Age (years) at onset of pain Number of clinicians co:~sulted

Mean 63.9 6.0 32.0 6.5

SD

23.3 5.0 11.0 3.7

mation as to the specialty of the clinicians was obtained as far as the fourteenth clinician consulted by a given patient. The percentage of nonprescribing clinicians stayed about the same. For example, on the fifth pass 91 clinicians were consulted including 2 chiropractors and 1 psychologist. It is apparent that more than 95 %0 of the many consultations patients underwent could have resulted in obtaining prescriptions for medication. The second labeling variable is the use of narcotic analgesics and sedatives/tranquilizers. Two measures were constructed. Patients were also asked about the medications they took during the previous year. Since narcotic analgesic and sedative/tranquilizer use is, in part, under the influence of the prescriber, it may provide a clue to the clinician's belief in the psychological origin of the pain. O f these two classes of drugs 10.6% (N = 16) of cases employed narcotic analgesics but no sedatives/tranquilizers, 25.8% ( N = 39) sedatives but no narcotic analgesics, 34.4% ( N = 52) both, and 29% (N = 44) neither. Two measures were constructed: one consisted of the percentage of cases who employed narcotic analgesics (45%, N = 68) and involved combining those who use both classes of drugs (34.4%) with those employing narcotic analgesics only (10.6%). We next combined those who took sedatives only (25.8%, N = 39) with those who took both cases of drugs (34.4%, N = 52) in order to calculate those who employed sedative/tranquilizer (60.3%, N = 91). We are concerned here with narcotic analgesics and tranquilizers as labeling variables, not as clinical variables. Thus, we permit overlap in the two categories since the research question has to do with the drugs as indicators of clinician's attitudes and not their inherent pharmacologic properties. Finally, the third labeling variable is taken f r o m the stigma questionnaire described below. Patients were asked whether they "have had doctors tell me that the pain I feel is all imaginary." A majority of cases (52.1%) agreed with this item. The correlation between this variable and the number of clinicians consulted for facial pain is .35 (p < .001). This shows that the number of clinicians consulted is related positively to the likelihood of having T M P D S labeled for the respondent as a disorder of psychological origin.

592

Marbach, Lennon, Link, and Dohrenwend Personality Measures

The personality variables included in this study were selected from a larger list because they differentiated these TMPDS patients from nonpatient controls (Marbach et al., 1988). Therefore, they could have a bearing on the second alternative hypothesis concerning personality problems among the patients. The first variable was locus of control orientation, which was measured by Rotter's Locus of Control Scale (Rotter, 1966). There is extensive empirical evidence that persons with an external locus of control expectancy have greater proneness to psychopathology than do those with an internal expectancy (reviewed by Lefcourt, 1981). The second measure was a 27-item measure of nonspecific distress or demoralization (Dohrenwend et al., 1980) derived from eight highly intercorrelated symptom scales measuring anxiety, sadness, poor self-esteem, hopelessness-helplessness, dread, confused thinking, psychophysiological symptoms, and perceived physical health. These measures of distress and locus of control expectancy were found to be stable. Using monthly measures of demoralization the correlation between month 1 and month 11 was .79. The correlation of the Rotter Locus of Control Scale measured at entrance into the study and 1 year later was .76. The present study used, as a measure of health or illness behavior, a 15-item scale of Illness Attitude that incorporates five scales developed by Kellner et al. (1983-1984). These scales, based on the work of Pilowsky (1967), were designed to measure psychopathology associated with hypochondriasis, which in turn can be responsible for abnormal illness behavior. The scales are worry about illness, hypochondriacal beliefs, fear of dying, disease phobia, and bodily preoccupation. TMPDS patients in this study had been found in a prior study to be more external in locus of control, to be more demoralized, and to show more abnormal illness behavior than controls (Marbach et al., 1988). Personality measures that did not show statistically significant case/control differences are not explored further in the present paper since patients do not show unusually high scores on them. Those measures excluded were denial as a way o f coping; sensation-seeking, which has been found to be related to substance abuse; and the Marlowe-Crowne Need for Approval Scale, a measure of social desirability response set. Stigma Scales

In order to measure perceived stigma, the Stigma of Facial Pain Questionnaire developed by Link, Marbach, and Lennon (available from the authors) was used. This questionnaire consisted of 25 Likert items, with responses ranging from strongly agree, coded "6," to strongly disagree, coded "1." Many of the items in the questionnaire were adapted from measures

Sources of Stigma for Facial Pain Patients

593

designed to assess the stigma associated with mental illness (Link, 1982, 1987; Link and Cullen, 1983). Other items were designed specifically for use with chronic pain patients and chosen after conducting open-ended interviews with TMPDS patients about their stigma experience. All of the items were pretested on a separate pool of patients to be certain they were clear, applicable, and salient to facial pain patients. Interviews were also conducted with these patients after they completed the questionnaire in order to identify and eliminate problematic items. Of the 151 patients who entered the study, 147 completed this questionnaire. The stigma items were factor analyzed using principal-components analysis and four orthogonal factors were extracted (Lennon et al., 1989). This paper focuses on two of the scales derived from this analysis. They are Estrangement and Attribution to Personality Problems, which are presented in Table II along with the percentage agreement for each of the items. The other two scales which measure the stigma management strategies of secrecy and disclosure are described elsewhere (Lennon et al., 1989).

Table IL Perceived Stigma Scales: Percentage Agreement, Item Means, and Standard Deviations Estrangement Scale 1. There is a part of me that only other people who have experienced facial pain can understand. 2. One problem with having facial pain is that people don't believe that you really hurt. 3. I have sometimes wished that people could see my pain. 4. Having facial pain has made me feel very different from other people. 5. Most people have no idea what it is like to have facial pain. 6. I often feel totally alone with my pain. Attribution to Personality Problems Scale 1. W h e n people learn that you have been treated for facial pain, they begin to search for flaws in your personality. 2. Many people assume that having facial pain is a sign of personal weakness. 3. Most people believe that a person with facial pain is just as emotionally stable as the average person. (R) + ~ 4. Many people assume that a person who suffers from facial pain has a deep seated psychological problem as well. 5. People have a way of associating the occurrence of facial pain with psychiatric difficulties.

% agree

Item mean

SD

76.4

4.9

1.3

57.9

4.4

1.6

70.5

4.6

1.6

39.6

3.6

1.7

81.0 67.9

5.1 4.5

1.1 1.6

19.6

2.7

1.6

15.1

2.4

1.6

46.4

3.8

1.6

29.7

3.2

1.6

37.4

3.5

1.6

a(R) + indicates that item scoring was reversed to calculate scale scores.

594

Marbach, Lennon, Link, and Dohrenwend

The Estrangement Scale captures feelings of alienation, isolation, being misunderstood, and being set apart from others by pain. Five of the six items are endorsed by the majority of patients. The invisibility and lack of objectification of one's pain as well as the inability to share the pain experience with others are the matters portrayed in the items describing the strong sense of estrangement among patients. The next five items in Table II constitute the Attribution to Personality Problems Scale. This scale reflects patients' beliefs about how others perceive or label TMPDS with regard to its potential psychological origins. The scale mirrors the expectation that others would respond in a rejecting or stigmatizing way if they were aware the person suffered from facial pain. A minority (19.6%) of patients accepts the generalization even in the most extreme item, that many others attribute facial pain to personality flaws or weakness. Reliabilities of the scales were assessed using Cronbach's alpha. For Estrangement, alpha = .84, and for Attribution, alpha = .82. The intercorrelation of the two scales is .52, indicating that while they have much in common, distinct aspects of perceived stigma are measured by each. Control Variables

In these analyses, the possible confounding influence of demographic differences are controlled statistically. The sociodemographic covariates included are a dummy variable representing whether the respondent works outside the home or not (coded "1" when the respondent is employed), years of education, and two dummy variables representing partner status: "never married" and "formerly married." The reference group (coded "0") consists of patients who are currently married or who have been living with a mate for a year or longer.

RESULTS The main research question is whether pej orative clinician labeling contributes to feelings of estrangement and beliefs about others' perception of personality flaws. Alternatively, it is possible that clinical factors such as the duration of TMPDS and severity of pain or underlying personality problems can explain these feelings and beliefs. The empirical investigation of these issues begins with presentation of bivariate associations between the stigma variables (estrangement and attribution) and the labeling variables. We then proceed to determine whether these associations remain when we control for the other possible determinants of stigma perceptions as described by our alternative hypotheses.

Sources of Stigma for Facial Pain Patients

595

Main Hypothesis Related to Clinician Labeling and Stigma Statistically significant bivariate associations were found between indicators of labeling and the stigma measures. Estrangement is related, as expected, to number of doctors consulted (r = .41, p < .001) and being told that the pain is imaginary (r = .53, p < .001). Moreover, as expected, the Attribution to Personality Problems Scale was significantly related to number of doctors consulted (r = .34, p < .001) and being told by a doctor that the pain is imaginary (r = .37, p < .001). Narcotic analgesic use but not tranquilizer use was positively related to estrangement (r = .28, p < .001) and attribution to personality problems (r = . 17, p < .05). These bivariate associations allow for the possibility that clinician labeling is a source of stigma among T M P D S patients. Alternative Hypothesis Related to Clinical Variables Nevertheless, the possibility still exists that relationship can be explained by the severity and duration of the pain itself. In order to investigate their alternative, bivariate associations between two pain variables and the stigma measures were examined. Severity of pain is significantly related to estrangement (r = .32, p < .001) but not to attribution to personality problems (r = .09, p > . 10). Duration of pain is significantly related to estrangement (r = .21, p < .05) and attribution to personality problems (r = .19, p < .05). This finding indicates that additional analyses are necessary in order to determine the role of labeling vs. severity and duration o f pain in stigma perception. Alternative Hypothesis Related to Personality Problems With regard to personality problems, demoralization is significantly related to estrangement (r = .42, p < .001) and attribution to personality problems (r = .19, p < .05). External locus of control is significantly related to attribution to personality problems (r = .18, p < .05) but not to estrangement (r = . 13, p > . 10). Illness attitudes were not significantly related to either stigma measure. These findings indicate that additional analyses are necessary in order to determine the role of personality factors vs. labeling in stigma perception. Independent Strength of Pain, Labeling, and Personality Problems on Stigma The data were analyzed using hierarchical multiple regression to assess the independent effects of labeling, pain duration, pain severity, and person-

596

Marbach, Lennon, Link, and Dohrenwend

ality problems, after controlling statistically for demographic factors. In this procedure, the coefficient of each variable represents its effect after all other variables in the equation are controlled statistically. The R 2 indicates the variance in the outcome explained by the set of independent variables in the equation. In order to form a context that will allow an assessment of the relative strength of labeling, clinical factors, and personality problems, we began by entering background characteristics as control variables into Equation A (see Tables III and IV for equations predicting estrangement and attribution to personality problems, respectively). These were years of education, marital status, and employment status. Next we entered pain variables (see Equation B). Since one of the main research questions asks whether clinician labeling explains perception of stigma above and beyond pain variables, we enter the pain variables before evaluating the effect of labeling. The pain variables included age at onset, duration, and severity of pain. Labeling variables were entered into Equations C and D. In Equation C we entered the number of doctors ever consulted for TMPDS and the use of narcotic analgesics and tranquilizers. Equation D includes the final labeling variable, whether a patient had been told her pain was imaginary. Equations E and F address the second alternative hypothesis that patients' perceptions of stigma are the result of personality problems. Equation A in Tables III and IV shows that the control variables account for small proportions of the variance in the Estrangement (6%) and Attribution to Personality Problems (4%) scales. When the pain variables are added (see Equation B), the amount of variance explained jumps a sizable 13% (to 19%) for estrangement and 4% (to 8%) for attribution to personality problems. For estrangement, both duration and severity are significant contributors to the explained variances; for attribution, duration is significant. When in Equation C, the labeling variables of number of clinicians and drug usage are added, while controlling statistically for the pain variables, these labeling variables account for a striking increase in the explained variance for both stigma scales. For estrangement, the percentage variance explained jumps a full 13~ indicating that labeling variables of number of clinicians and drug use uniquely account for a substantial amount of variability in this outcome. For attribution, fully 11% of the variance is uniquely attributable to these variables, more than twice the amount explained by the other variables in the model. However, only number of clinicians consulted appears to be strongly related to estrangement; drug use does not show a significant relation. For attribution as well, the number of doctors consulted shows an independent effect.

**p _< .01. ***p __..001.

*p < .05.

R2

Education Formerly married/ mated Never married/mated Employment status Age at onset Duration of pain Pain severity Number of doctors consulted Narcotic use Tranquilizer use Doctor said pain is imaginary Demoralization Rotter locus of control Illness attitude

(.260) (.227) (.238)

-.402

.06

(.041)

-.062

-.399 -.121

(SE)

b

A

.003* .013"

-.014

-.340 -.268 -.164

-.056

b

B

.19

(.243) (.230) (.231) (.009) (.002) (.004)

(.039)

(SE)

.32

(.024) (.181) (.184)

.092*** .371 .071

(.037)

(SE)

(.232) (.218) (.222) (.008) (.001) (.004)

C

.333 -.156 .006 -.014 .002 .010"*

-.053

b

.45

(.042)

.239***

(.209) (.196) (.201) (.007) (.001) (.003)

(.033)

(SE)

(.022) (.163) (.165)

D

.054* .395* .029

-.226 -.099 -.147 -.016" .002 .005

-.045

b

(.042) (.123) (.416)

.050* .354* .023 .221"** .295* .338

.51

(.022) (.161) (.165)

.002 .004

-.017

(.207) (.195) (.203) (.008) (.001) (.004)

-.199 -.184 .029

(SE) (.033)

E -.037

b

.203*** .277 .270 .142

.046* .387* .032

.50

(.047) (.145) (.462) (.157)

(.024) (.172) (.180)

(.221) (.207) (.227) (.008) (.001) (.004)

- .043 -.178 .123 .088 -.015 .002 .005

(SE) (.037)

b

F

Table III. Regression of Estrangement on Demographic (A), Clinical (B), Labeling (C and D), Personality (E), and Illness Attitude (F) Variables

*p _< ,05, **p _ .01. ***p _< .001.

R2

Education Formerly married/ mated Never married/mated Employment status Age at onset Duration of pain Pain severity Number of doctors consulted Narcotic use Tranquilizer use Doctor said pain is imaginary Demoralization Rotter locus of control Illness attitude

.04

-.443 -.277 .475 .011 .004* .004

(.310) (.271) (.284)

-.501 .358

-.464

b

-.020

(SE)

(.049)

b

-.015

A

B (SE)

.08

(.307) (.290) (.292) (.011) (.002) (.005)

(.049)

b

(SE)

(.031) (.234) (.237)

.104" .323 .194

.19

(.299) (.282) (.285) (.011) (.002) (.005)

(.047)

-.445 -.160 .653* .010 .003 .000

-.018

C b

(SE)

(.059)

.185'*

.25

(.031) (.226) (.229)

(.290) (.272) (.280) (.010) (.002) (.005)

(.046)

.074* .342 .161

-.362 -.116 .535 .009 .003 -.003

-.012

D b

.172"* .074 1.247"

.073* .296 .200

-.236 -.183 .685 .007 .002 -.001

-.018

E (SE)

.29

(.060) (.173) (.585)

(.031) (.227) (.232)

(.292) (.275) (.285) (.011) (.002) (.005)

(.047)

b

.127" .032 1.238" .106

.074* .363 .215

-.155 -.090 .817" .010 .002 .001

-.010

F

.29

(.064) (.200) (.636) (.216)

(.033) (.237) (.248)

(.304) (.285) (.312) (.01t) (.002) (.005)

(.050)

(SE)

Table IV. Regression of Attribution to Personality Problems on Demographic (A), Clinical (B), Labeling (C and D), Personality (E), and Illness Attitude (F) Variables

Sources of Stigma for Facial Pain Patients

599

In Equation D we add the critical variable of being told by clinicians that one's pain is imaginary. As predicted, this variable is strongly associated with both stigma variables, uniquely accounting for 13% of the variance in estrangement and 6% of the variance in attribution. Furthermore, adding this variable helps explain why the number of clinicians variable was strongly associated with the stigma variables in the bivariate analyses. Note that in Equation D the size of the regression coefficient for number of clinicians declines by 40% (from .09 to .05) for the estrangement scale and 30% (from. 10 to .07) for the attribution scale. In addition, for the estrangement scale there is a precipitous drop in the significance level, from .001 to .05. This suggests that a major reason that contacting many clinicians was associated with stigma feelings in the bivariate analyses is that such contact puts one at high risk of being exposed to the clinician sanctioned idea that the pain is due to personality problems. When demoralization and locus of control are added, they account for a small increase in the proportion of the variance in both the Estrangement (6%) and Attribution to Personality Problems (4%) scales (see Equation E). For estrangement, demoralization is significant (p < .05) but has little effect on the strength of the labeling variables. Locus of control shows a significant independent effect for attribution. These findings suggest that perceived stigma shows only a small direct relationship to these personality measures. The illness attitude scale which measures many concepts subsumed under the rubric of abnormal illness behavior does not contribute to the overall variance (Equation F), despite the strong case/control difference on this variable (Marbach et al., 1988). However, the addition of the illness attitude scale attenuated the contribution of demoralization to estrangement (p < .06) (see Table III, Equation F compared to E). For attribution, attenuation also took place for the imaginary item (p _< .05) after introducing illness attitudes into the equation.

DISCUSSION The purpose of this study was to determine whether TMPDS patients feel stigmatized because of clinician labeling. The results show that when clinicians are judged by patients to attribute the origins of TMPDS to psychological causes, patients feel stigmatized. The stigma perceived by TMPDS patients cannot be explained by the severity or duration of their pain or by the personality factors investigated. There are several limitations of the study to consider before we discuss the implications of these findings. One of these stems from the fact that this

600

Marbach, Lennon, Link, and Dohrenwend

is a study of patients' perceptions. Future investigations could expand on these findings by assessing directly how clinicians explain TMPDS to their patients. Such reports would give us insight into whether clinicians' views of TMPDS patients are accurately reported by the patients or are the result of misunderstood communications. In addition to the information given about clinician-patient encounters, clinician reports would help overcome problems of potentially inflated correlations due to shared method variance that arise from using the same method to collect information on independent and dependent variables (Campbell and Fiske, 1959). For example, having been told that the pain is imaginary and the scales of perceived stigma are based on patients' responses to the same questionnaire. The addition of reports from clinicians as to whether they see the pain as imaginary would help in dealing with possible problems of shared method variance. The relationship between narcotic use for the relief of chronic pain and stigma also requires further investigation. Perhaps the transaction necessary to obtain narcotics may in itself be stigmatizing even though they are prescribed by a clinician (Portenoy and Foley, 1986; Sternbach, 1987). It is also possible that those who use narcotics may feel estranged as a result of the dysphoric action of these drugs. The larger amount of explained variance of labeling and stigma to the Estrangement scale compared to the Attribution to Personality Problems scale requires comment. The estrangement items describe the feelings and perceptions of the sufferer. The attribution items attempt to capture feelings about how patients believe others view them. It is possible that this scale does not measure the full range, the variety, and some of the subtleties of the respondents' beliefs about the views of others. More work may be warranted on the measurement of this aspect of perceived stigma. We must also insert a caveat about ambiguity in the direction of relationships among the variables investigated in this cross sectional study. For example, people who feel stigmatized may be sensitive to real or imagined rejection by clinicians and therefore change from one clinician to another frequently. This, and not clinical results, may account for why TMPDS patients consult so many clinicians. The causal direction of the associations cannot be determined with our data and for this reason reverse or reciprocal causation has to be considered a possibility. Longitudinal studies of cohorts of TMPDS cases who seek treatment for the first time would be necessary in order to answer the question conclusively. Meanwhile, our results are, for the most part, consistent with the following explanation of how feelings of being stigmatized for TMPDS develop. Care seeking can be viewed, in Kotarba's (1983) terms, not only as a search for healing but as a search for meaning. Confronted with chronic unexplained pain, afflicted individuals quite naturally seek reasons for and relief from the condition that torments them. Their search inevitably takes them to those

Sources of Stigma for Facial Pain Patients

601

our culture legitimizes as healers. Unfortunately, due to the condition's unexplained nature and to the limited repertory of truly effective palliatives, the healers they contact are unable to provide credible explanations for the pain. As the results concerning the number of clinicians contacted shows, an extended search for effective treatments and satisfactory explanations ensues. Clinicians, unable to explain a condition in biomedical terms, especially since traditional biomedical treatments frequently fail, attribute the disorder as well as the treatment failure to psychological factors. When clinicians communicate this, patients develop beliefs about being perceived as suffering from flawed personalities and consequent feelings of being estranged. In brief, patients feel stigmatized. Our findings suggest that efforts should be made by clinicians to reduce patients' feelings of estrangement. However, this may be problematic for chronic facial pain patients given the evidence that clinicians may come to resent chronically ill patients for whom they play a caretaking role (Maslach and Jackson, 1981; Jackson and Maslach, 1982). It is suggested that among other attitudes, clinicians tend to depersonalize those they care for who do not improve over a long period of time, increasing feelings of estrangement, rejection, and stigma. It is an important goal to make clinicians aware of how their words and deeds contribute to patients' estrangement and feeling flawed, and the potential health consequences. For example, we have demonstrated that, in TMPDS cases, perceptions of low self-esteem and feeling helpless and hopeless are significantly correlated with decreased mitogen-induced lymphocyte proliferative responses (Marbach et al., 1990b). To the extent that these perceptions are in part the result of feeling stigmatized, it is tempting to suggest that a cognitive psychological state that leads from stigma to low self-esteem and helplessness could result in alteration in immune function. This in turn could predispose to the increased physical illness found in TMPDS cases (Marbach et al., 1988). This is, however, conjecture. Clinicians once sensitized to patients' needs not to lose face or be stigmatized can take appropriate steps. Therefore, we have a clear-cut indication for a treatment strategy that moderates the sources of estrangement, namely, reducing or eliminating stigmatizing labeling. Future research will determine whether saving face, by and for the patient, turns out to be tantamount to promoting good physical as well as psychological health.

ACKNOWLEDGMENTS

This research was supported by NIH Grant DE05989. We wish to thank Drs. Karen Raphael and Marylene Cloitre and the anonymous reviewers for valuable comments on early drafts of this paper and James Sandy for assistance with data analysis.

602

Marbach, Lennon, Link, and Dohrenwend REFERENCES

Ahlers, T. A., Yunus, M. B., Riley, S. D., Bradley, J. M., and Masi, A. T. (1984). Psychological factors associated with primary fibromyalgia syndrome. Arth. Rheum. 27:1101-1106. American Psychiatric Association (1987). Diagnostic and Statistical Manual o f Mental Disorders (DSM-III-R), American Psychiatric Association, Washington, D.C. Angell, M. (1985). Disease as a reflection of the psyche. N. Engl. J. Meal. 312: 1570-1572. Campbell, D. T., and Fiske, D. W. (1959). Conversent and discriminant validation by the multitrait-multimethod matrix. Psych. Bull. 56: 81-105. Cassell, E. J. (1982). The nature of suffering and the goals of medicine. N. Engl. J. Med. 306: 639-645. Clark, G. T. (1984). A critical evaluation of orthopedic interocclusal appliance therapy: Effectiveness for specific symptoms. J. Am. Dent. Assoc. 108: 364-368. Dohrenwend, B. P., Shrout, P. E., Egri, G., and Mendelsohn, F. S. (1980). Nonspecific psychological distres and other dimensions of psychopathology. Arch. Gen. Psychiat. 37: 1229-1236. Eversole, L. R., Stone, C. E., Matheson, D., and Kaplan, H. (1985). Psychometric profiles and facial pain. Oral Surg. Oral Med. Oral Pathol. 60: 269-274. Feinmann, C. (1983). Psychogenic facial pain: Presentation and treatment. J. Psychosom. Res. 27: 403-410. Ford, A. B. (1962). Reaction of physicians and medical students to chronic illness. J. Chron. Dis. 15: 785-794. Ford, C. (1983). The Somatizing Disorders: Illness as a Way o f Life, Elsevier, New York. Goffman, E. (1963). Stigma: Notes on the Management o f Spoiled ldentity, Prentice-Hall, Englewood Cliffs, N.J. Greene, C. S., and Marbach, J. J. (1982). Epidemiologic studies of mandibular dysfunction: A critical review. J. Prosthet. Dent. 48: 184-190. Heloe, B., and Heloe, L. A. (1975). Characteristics of a group pain patients with temporomandibular joint dysfunction disorders. Commun. Dent. Oral Epidemiol. 3: 72-79. Hilbert, R. A. (1984). The accultural dimensions of chronic pain: Flawed construction and the problem of meaning. Soc. Problems 31: 365-378. Hollingshead, A. B., and Redlich, F. C. (1958). Social Class and Mental Illness: A Community Study, Wiley, New York. Huskisson, E. C. (1983). Visual analogue scales. In Melzack, R. (ed.), Pain Measurement and Assessment, Raven Press, New York, pp. 33-37. Illich, I. (1976). Medical Nemesis: The Expropriation o f Health, Random House, New York. Jackson, S. E., and Maslach, C. (1982). After-effects of job-related stress: Families as victims. J. Occupat. Behav. 3: 63-77. Jones, E. E., Farina, A., Hastorf, A. H., etal. (1984). SocialStigma, W. H. Freeman, New York. Kellner, R., Abbott, P., Pathak, D., Winslow, W. W., and Umland, B. E. (1983-1984). Hypochondriacal beliefs and attitudes in family practice and psychiatric patients. Int. J. Psychiat. Med. 13: 127-139. Kleinknecht, R. A., Mahoney, E. R., and Alexander, L. D. (1987). Psychosocial and demographic correlates of temporomandibular disorders and related symptoms: An assessment of community and clinical findings. Pain 29: 313-324. Kotarba, J. A. (1983). Chronic Pain: Its Social Dimensions, Sage, Beverly Hills, Calif. Laskin, D. M. (1969). Etiology of the pain dysfunction syndrome. J. Am. Dent. Assoc. 9: 147-153. Lefcourt, H. M. (1981). Locus of control and stressful life events. In Dohrenwend, B. S., and Dohrenwend, B. P. (eds.), StressfulLife Events and Their Context, Prodist, New York, pp. 157-166. Lennon, M. C., Link, B. G., Marbach, J. J., and Dohrenwend, B. P. (1989). The stigma of chronic facial pain and its impact on social relationships. Soc. Problems 36:117-134. Link, B. G. (1982). Mental patient status, work, and income: An examination of the effects of a psychiatric label. Am. Sociol. Rev. 47: 202-215. Link, B. G. (1987). Understanding labeling effects in the area of mental disorders: An assessment of the effects of expectations of rejection. Am. Sociol. Rev. 52: 96-112.

Sources of Stigma for Facial Pain Patients

603

Link, B. G., and Cullen, F. T. (1983). Reconsidering the social rejection of ex-mental patients: Levels of attitudinal response. Am. J. Comm. Psychol. 11: 261-273. Lipton, J. A., and Marbach, J. J. (1983). Components of the response to pain and variables influencing the response in three groups of facial pain patients. Pain 16: 343-359. Lipton, J. A., and Marbach, J. J. (1984a). Predictors of treatment outcome in patients with myofascial pain-dysfunction syndrome and organic temporomandibular joint disorders. J. Prosthet. Dent. 51: 387-393. Lipton, J. A., and Marbach, J. J. (1984b). Ethnicity and the pain experience. Soc. Sci. Med. 19: 1279-1298. Marbach, J. J., and Lipton, J. A. (1978). Aspects of illness behavior in patients with facial pain. J. Am. Dent. Assoc. 96: 630-638. Marbach, J. J., and Lipton, J. A. (1987). Biopsychosocial factors of the temporomandibular pain dysfunction syndrome: Relevance to restorative dentistry. Dent. Clin. 1~: Am. 31: 473-485. Marbach, J. J., Lipton, J. A., Lund, P. B., Delahanty, F., and Blank, R. T. (1978). Facial pains and anxiety levels: Considerations for treatment. J. Prosthet. Dent. 40: 434-437. Marbach, J. J., Lennon, M. C., and Dohrenwend, B. P. (1988). Candidate risk factors for temporomandibular pain and dysfunction syndrome: Psychosocial, health behavior, physical illness and injury. Pain 34: 139-151. Marbach, J. J., Raphael, K. G., Dohrenwend, B. P., and Lennon, M. C. (1990a). Etiology of TMPDS revisited: The validity of tooth grinding measures. J. Am. Dent. Assoc. 120: 327-333. Marbach, J. J., Schleifer, S. J., and Keller, S. E. (1990b). Facial pain, distress, and immune function. Brain Behav. Imrnun. (in press). Maslach, C., and Jackson, S. E. (1981). The measurement of experienced burnout. J. Occupat. Behav. 2: 99-113. Mechanic, D. (1980). The experiences and reporting of common physical complaints. J. HIth. Soc. Behav. 21: 146-155. Melzack, R. (1975). The McGill Pain Questionnaire: Major properties and scoring methods. Pain 1: 277-299. Merskey, H. (1986). International Association for the Study of Pain. Classification of chronic pain. Pain Supp., 3: 559-560. Merskey, H. (1988). Regional pain is rarely hysterical. Arch. Neurol. 45: 915-918. Nuprin (1985). Nuprin Pain Report, Sept. Louis Harris and Associates, New York, p. 102. Osterweis, M., Kleinman, A., and Mechanic, D. (1987). Pain and Disability: Clinical, Behavioral and Public Policy Perspectives, National Academy Press, Washington, D.C. Pearlin, L. I. (1989). The sociological study of stress. J. filth. Soc. Behav. 30: 241-256. Pilowsky, I. (1967). Dimensions of hypochondriasis. Br. J. Psychiat. 113: 89-93. Portenoy, R. K., and Foley, K. M. (1986). Chronic use of opioid analgesics in non-malignant pain: Report of 38 cases. Pain 25: 171-186. Ready, L. B., Sarkis, E., and Turner, J. A. (1982). Self-reported vs. actual use of medications in chronic pain patients. Pain 12: 285-294. Rotter, J. B. (1966), Generalized expectancies of internal versus external control of reinforcement. Psychol. Monogr. 80 (609). Rudy, T. E., Turk, D. C., Zaki, H. S., and Curtin, H. B. (1989). An empirical toxometric alternative to traditional classification of temporomandibular disorders. Pain 36:311-320. Rugh, J. D., and Harlan, J. (1988). Nocturnal bruxism and temporomandibular disorders. In Jankovic, J., and Tolosa, E. (eds.), Advances in Neurology: Facial Dyskinesias, Raven Press, New York, Vol. 49, pp. 329-341. Salter, M., Brooke, R. I., Merskey, H., Fichter, G. F., and Kapasianyk, D. H. (1983). Is the temporo-mandibular pain and dysfunction syndrome a disorder of the mind? Pain 17: 151-166. Schnurr, R. F., Brooke, R. I., and Rollman, G. B. (1990). Psychosocial correlates of temporomandibular joint pain and dysfunction. Pain 42: 153-166. Speculand, B., Hughes, A. O., and Goss, A. N. (1984). Role of recent stressful life events experiences in the onset of TMJ dysfunction pain. Comrnun. Dent. Oral Epidemiol. 12: 197-202.

604

Marbach, Lennon, Link, and Dohrenwend

Sternbach, R. A. (1987). Letter. Pain 29: 257-258. Sternbach, R. A., Murphy, R. W., Timmermans, G., Greehoot, J. H., and Akeson, W. H. (1974). Measuring the severity of clinical pain. In Bonica, J. J. (ed.), Advances in Neurology, Vol. 4, Raven Press, New York, pp. 281-288. Swanson, D. W., Maruta, T., and Wolff, V. A. (1986). Ancient pain. Pain 25: 383-387. Thoits, P. A. (1986). Multiple identities: Examining sexual and marital status differences in distress. Am. Sociol. Rev. 51: 259-272. Tursky, B. (1974). The development of pain perception profile: A psychophysical approach. In Weisenberg, M., and Tursky, B. (eds.), Pain: New Perspectives in Therapy and Research, Plenum Press, New York, pp. 281-288. Volinn, I. J. (1983). Health professionals as stigmatizers and destigmatizers of diseases: Alcoholism and leprosy as examples. Soc. Sci. Med. 17: 385-393. Von Korff, M., Dworkin, S. F., Le Resche, L., and Kruger, A. (1988). An epidemiologic comparison of pain complaints. Pain 32: 173-184. Wallenstein, S. L., Heidrich, G., Kaiko, R., and Houde, R. W. (1980). Clinical evaluation of mild analgesics: The measurement of clinical pain. Br. J. Clin Pharmac. 10:319S-327S.