~ Pergamon S (96)

Transcription

1 ~ Pergamon S (96) Behav. Res. Ther. Vol. 34, No. 9, pp , 1996 Copyright 1996 Elsevier Science Ltd Printed in Great Britain. All rights reserved /96 $ DEVELOPMENT OF THE BODY DYSMORPHIC DISORDER EXAMINATION JAMES C. ROSEN* and JEFF REITER University of Vermont, Burlington, VT 05405, U.S.A. (Received 31 May 1995, revised 28 February 1996) Summary--The Body Dysmorphic Disorder Examination (BDDE) is a semi-structured clinical interview designed to diagnose body dysmorphic disorder and to measure symptoms of severely negative body image. It taps into preoccupation with and negative evaluation of appearance, self-consciousness and embarrassment, excessive importance given to appearance in self-evaluation, avoidance of activities, body camouflaging, and body checking. The BDDE had adequate internal consistency and test-retest and interrater reliability. It correlated with measures of body image, negative self-esteem, and psychological symptoms, and was sensitive to change following treatment of body dysmorphic disorder. The BDDE distinguished body dysmorphic disorder patients from clinical and non-clinical control subjects and agreed with other clinicians' diagnosis of body dysmorphic disorder. The BDDE provided unique information in predicting clinical status when controlling for psychological adjustment and other measures of body image. Copyright 1996 Elsevier Science Ltd INTRODUCTION Body dysmorphic disorder (BDD) is a distressing and disabling body image disorder that has been described in psychiatric literature for over a century (Morselli, 1886). According to the diagnostic criteria (American Psychiatric Association, 1987), the essence of BDD is a preoccupation or excessive concern with a non-existent or slight physical defect in a normal appearing person. Patients with BDD can complain of nearly any aspect of their appearance, from localized, specific defects to vague feelings of ugliness. The most common features are head and body hair, facial features, skin blemishes, thighs, stomach, breasts, and buttocks (Phillips, McElroy, Keck, Pope & Hudson, 1993; Rosen, 1995). According to the diagnostic criterion added to the DSM-IV (American Psychiatric Association, 1994), the preoccupation causes significant impairment in social functioning or marked distress. The prevalence of BDD is unknown, but it is probably more common than generally believed because most persons with BDD do not seek treatment (Phillips, 1991). Recently it has been receiving more attention, but a major problem in studying and treating body dysmorphic disorder is that there is no validated standard measure or diagnostic schedule for BDD. That is the gap the present project was designed to fill. Although distinct psychological and biological factors may play a role in BDD, the body image complaints of BDD patients can overlap with other types of appearance concerns. Therefore, an adequate measure of body dysmorphic disorder must distinguish it from normal body dissatisfaction and from body image problems in other clinical populations (e.g. persons with real physical deformities). To accomplish this, we believe the measure must inquire about feelings of shame in social situations, excessive importance given to physical appearance in self-evaluation, and body checking and impairment of activities. Moreover, the measure should supplement the diagnostic criteria, which presently are vague and highly subjective, by including more of the typical BDD symptoms, standardized severity ratings, and an operational definition of BDD. The content of the measure should not be gender biased or limited to certain types of appearance complaints. Presently, no published measure of body image meets these requirements. Most popular body image measures were designed for use with eating disorders and, consequently, the content refers mainly to women's weight and body shape dissatisfaction.* Other popular measures of body image that are appropriate for men and women with any type of appearance complaint do not tap into all the clinical features of BDD.t *Author for correspondence; address: j_rosen@dewey.uvm.edu. Please enclose $3.00 for the BDDE. BRT 34/9 r 755

2 756 James C. Rosen and Jeff Reiter To date, the use of formal assessment techniques with BDD patients is very limited. The Yale-Brown Obsessive Compulsive Scale (Y-BOCS) was modified by Phillips (1993) to measure obsessions and compulsions in BDD. Among other changes, the phrase 'thoughts about body defect' was inserted into the standard questions. The scale taps into the distressing thoughts about the defect and efforts made to control them, but it is limited to obsessive compulsive type symptomatology in BDD patients. BDD patients score lower on the measure after therapy (Neziroglu & Yaryura-Tobias, 1993a), but no formal psychometric studies of the modified Y-BOCS have been conducted. Neziroglu and Yaryura-Tobias (1993b) described the Overvalued Ideation Scale with which the patient rates the strength of his or her belief in the defect. The authors state that this is a measure of insight. BDD patients report a high level of conviction on this scale. No psychometric studies are available for this scale. Only two reports used a traditional body image assessment technique. Both used figure drawings to measure body part size distortion: distortion of nose size in rhinoplasty patients (Jerome, 1992) and distortion of penis size in koro patients (Chowdhury, 1989). However, the Ss in these studies were not clearly diagnosed as BDD. The Body Dysmorphic Diagnostic Module (Phillips, 1994) is a checklist of six questions that correspond to the DSM-IV criteria. It assists the interviewer in making the diagnosis, but does not provide a quantitative measure of symptom severity. No psychometric studies are reported. The measure we developed is the Body Dysmorphic Disorder Examination (BDDE). It is a 34-item semi-structured clinical interview that usually requires min to administer. The BDDE may be too time-consuming to be used as a screening instrument. However, we designed it to not only diagnose BDD but also to give the clinician a detailed assessment of typical symptoms that could be targeted in treatment. We chose a clinician administration to help obtain accurate ratings, especially on items that require the S to make subtle distinctions between types of body image beliefs (e.g. believing that other people evaluate their defect negatively vs evaluate their character negatively because of the defect). The purpose of this paper is to present the development and content of the BDDE and studies on its reliability and validity. STUDY 1: DEVELOPMENT OF THEBDDE Item development We compiled all the body image complaints and associated features of BDD patients that were reported in the published case studies (Phillips, 1991) and then constructed a non-redundant list of symptoms. Items were written in the form of questions that would probe for the presence and severity of each feature. In a pilot study, the resulting interview was administered to a convenience sample of 22 clinical and non-clinical Ss, it was revised for clarity, administered to another convenience sample of 23 Ss, and revised again based on their feedback. The third version of the interview was given to a panel of 10 expert clinicians (doctoral psychologists and psychiatrists with diagnostic and treatment experience with BDD) to rate the content. They were provided with the DSM criteria and text for BDD (American Psychiatric Association, 1994) and asked to rate each item on the BDDE as appropriate or inappropriate for the assessment of BDD. Items that were not endorsed positively by at least 8 out of 10 raters were dropped. The resulting fourth version (version 3.2) of the BDDE was used in the studies that follow (see Appendix for a brief description of the items). Content The interview begins by asking the S to describe any aspect of physical appearance that he or she has disliked during the past four weeks (Item 1, Appendix). The perceived physical defect is rated by the interviewer as either 'not observable', 'observable but minimally defective---not abnormal', or 'definitely abnormal' (Item 2). Twenty-eight other items concern the severity of BDD *For example, the Eating Disorder Inventory (Garner, 1991), Body Shape Questionnaire (Cooper, Taylor, Cooper & Fairburn, 1987), Eating Disorder Examination (Cooper & Fairburn, 1987), and the discrepancy score between current and ideal body figure drawings (e.g. Williamson, Davis, Bennett, Gorenczny & Gleaves, 1989). tfor example, the Multidimensional Body-Self Relations Questionnaire (Brown, Cash & Mikulka, 1990) and measures of body part satisfaction (e.g. the Body Esteem Scale: Franzoi & Shields, 1984).

3 Body Dysmorphic Disorder Examination 757 Table 1. Recommended BDDE items for diagnosis of BDD Criterion A: Preoccupation with an imagined defect in appearance. If a slight physical anomaly is present, the person's concern is markedly excessive. No. 2 = 0 or 1 (defect not observable or minimal) No. 3 = 0 (somatic complaints concern appearance) No. 9 = 4 or higher (upsetting preoccupation) No. 10 or 11 = 4 or higher (self-consciousness and embarrassment) No. 18 = 4 or higher (importance of appearance) No. 19 = 4 or higher (negative self-evaluation due to appearance) Criterion B: The preoccupation causes clinically significant distress or impairment in social, occupational, or other important areas of functioning. No. 9 = 4 or higher (upsetting preoccupation) No. 10 or 11 = 4 or higher (self-consciousness and embarrassment) No. 13 = 4 or higher (distress when people notice appearance) all three of the above or: No. 23, 24, 25 or 26 = 4 or higher (avoidance of activity due to appearance) Criterion C: The preoccupation is not better accounted for by another mental disorder No. 34 = 0 (preoccupation with defect not better accounted for by another disorder) Note: See Appendix for description of items. symptoms over the past four weeks. The content covers six areas: (a) preoccupation with and negative evaluation of appearance, (b) self-consciousness, embarrassment, and feeling scrutinized in public, (c) excessive importance given to appearance in self-evaluation, (d) avoidance of social situations or activities in public and avoidance of physical contact with others, (e) camouflaging appearance with styles of dress, make-up, grooming, or body posture, and (f) body checking behavior in the form of self-inspection, repetitive grooming, reassurance seeking, and comparing self to others. Each BDD feature is probed with a standard question. Optional follow-up questions are suggested to clarify the S's response. Instructions for rating the S's response are provided for the interviewer. The 28 items are rated on a 0-6 scale, 0 indicating the feature is absent or did not occur in the past 28 days. If present, ratings of 1-6 represent the frequency of the feature (number of days present) or intensity (slight to extreme). The S is provided with a rating scale for each question and, in addition to answering the question in detail, is asked to suggest a rating. The interviewer takes this into consideration when making the final rating (see Appendix for endpoints on the scales). Several items are provided for diagnostic and information purposes only and are not included in a total severity score. First, the interviewer indicates if the complaint is exclusively related to another disorder, for example, complaints of being too fat in a S who has an eating disorder (Item 34). This rating is provided for the diagnostician's convenience, but further assessment might be required if the examiner suspects an eating disorder. Second, the S is asked to check from a list of beauty or corrective remedies (e.g. cosmetic surgery) that have been tried at any time (Item 33). Third, a rating is provided for the examiner to indicate if the patient complains exclusively of a somatic preoccupation that concerns bodily functions or sensations other than physical appearance (e.g. infestation with insects or emitting a foul odor; Item 3). If present, the examiner would rule out BDD and have to consider diagnosing a somatic delusional disorder, using further assessment. Finally, several questions are provided to determine the extent to which the S is convinced the defect truly exists or is significant (Item 22). This allows for a rating of insight and the possibility of a conviction held with delusional intensity.* Although the diagnosis of BDD is not conditional on the presence or absence of delusional thinking, these questions are provided for clinical information and to assist the interviewer in formulating an additional diagnosis of somatic delusional disorder. Scoring Diagnosis of BDD. The recommended criteria for the diagnosis of BDD using individual items from the Body Dysmorphic Disorder Examination is presented in Table 1. The following steps were *A criterion in the DSM-III-R was that the somatic preoccupation not be of delusional intensity, but this was omitted in the DSM-IV based on evidence that BDD patients can exhibit delusional thinking (Phillips & McElroy, 1993). The definition and importance of delusions in BDD are uncertain (de Leon, Bott & Simpson, 1989). However, because the DSM-IV specifies that BDD patients can receive the additional diagnosis of delusional disorder somatic type, we included ratings related to delusional thinking.

4 758 James C. Rosen and Jeff Reiter Table 2. Sample characteristics Clinical Non-clinical BDD Non-BDD University undergrads University staff Number (total 694): Men Women Total Age: Men (7.7) 33.1 ~ (10.7) 19.0 b (2.5) 38.9` (9.8) Women 34.5 ~ (10.3) 36.8" (11.4) 18.5 b (I.9) 41.9' (10.0) Education: (yr) Men 12.8 (I.0) 12.9" (3.7) 13.4 i (.7) 17.2 b (3.2) Women 15.2 ~b (2.0) 14.7 b (2.4) 13.3 c (.7) 15.3' (2.7) Marital status (men and women): Single 49% 52% 100% 26% Married 32% 31% 55% Divorced 19% 17% 19% Body Dysmorphic Disorder Examination: Men (total score) 103.9` (20.5) 63.7 b (23.3) 24.3' (15.8) (I 1.3) Women (total score) 90.7 i (15.0) 73.5 b (26.2) 42.2 c (24.4) (21.4) Note: Means are shown with standard deviations in parentheses. Row means with different superscripts are significantly different, P < 0.05 according to Student Newman Keuls post-hoc tests. taken to develop this set of cutoff scores. The 10 expert raters categorized each of the 34 BDDE items to criterion A, B or C of the DSM-IV (American Psychiatric Association, 1994) or none. Items with agreement on the categorization by at least eight raters were used in the next step. The raters were asked to pick a minimum number of items that were closest to the BDD criteria in the DSM. This resulted in 13 items that were recommended by at least eight raters. Some BDDE items were categorized as being relevant to both DSM criteria A and B. (See Table 1 for the items and scheme for diagnosis of BDD.) The cut score for most of the items was set at a minimum of 4. This indicates the feature was present on half to all days or was moderately severe to severe in intensity. The cut score was at least 1.96 SDs higher than normal, according to item means from the non-clinical university staff sample (see below). We considered setting a minimum total score as another criterion for the operational definition of BDD. However, in examining the BDDE scores of the clinical sample (see below), we found that no S who met all the criteria in Table 1 had a total score of less than 66, 1.78 and 4.46 SDs above the female and male university staff means. Therefore, it appears unlikely that a S who meets these criteria would have a normal total score. On the other hand, it was possible for a S to achieve a high total score (in excess of 66) without being significantly symptomatic on the essential BDD items. Therefore, a cutoff based solely on the total score is not recommended. Total symptom severity. The total score on the BDDE is calculated as the sum of ratings for the 28 symptom items (all the items except 1, 2, 3, 22, 33 and 34; 0 to 168 possible). This provides an overall severity index, taking into account all BDD symptoms, whether or not the S can be diagnosed as BDD. STUDY2: BASIC PSYCHOMETRIC PROPERTIES Subjects Method Subjects were 694 men and women from two clinical and two non-clinical samples (Table 2). The clinical Ss were 259 consecutive referrals to an outpatient clinic for assessment or treatment of BDD and other body image complaints. Referring professionals included mental health therapists, general physicians, plastic surgeons, orthopedists, dermatologists, and dieticians. Subjects were also recruited with newspaper announcements of a 'body image therapy' program. Common to all Ss was a primary referral concerning preoccupation with physical appearance. In addition to the battery of measures described below, Ss were examined independently by a psychologist or psychiatrist to identify cases of body dysmorphic disorder. The diagnosis of BDD was made using the Body Dysmorphic Disorder Diagnostic Module (Phillips, 1994), a brief interview using the SCID format (Structured Clinical Interview for DSM-IV). The independent examiners were familiar with the disorder and the reliability of the diagnostic module has been reported as adequate (Hollander, Cohen & Simeon, 1993). Subsequently, Ss were separated into

5 Body Dysmorphic Disorder Examination 759 two groups: BDD and Non-BDD. (The diagnostician was not provided with the results of the research measures. Subjects who presented initially with an obvious physical abnormality were not interviewed.) Body Dysmorphic Disorder Ss met the criteria for BDD according to the diagnostic interview (Ss were diagnosed by the independent clinician, not by the BDDE). Non-Body Dysmorphic Disorder Ss presented body image complaints but did not qualify as BDD: (a) 36. 1% of Non-BDD Ss had a real physical defect (30.5% were severely obese; 5.6% had abnormal facial features, skin abnormalities, or amputated limbs); (b) 6.2% of Ss had appearance complaints that were related to an eating disorder in partial remission [eating disorders were diagnosed with the Eating Disorder Examination (Cooper & Fairburn, 1987)]; and (c) 57.7% of Ss presented appearance preoccupation that was below clinical threshold for BDD. Of the non-clinical Ss, one control group consisted of 295 male and female university undergraduates. These were students from introductory psychology classes who were invited to participate in a study of 'gender and self-image' in exchange for extra course credit. The other non-clinical control group consisted of 140 male and female university staff. Potential Ss for this group were randomly selected from the university telephone directory and asked to participate in a study of 'self-image'. Sixty-nine percent of those contacted said they would participate and 66% actually completed the study. All volunteers were included; no Ss were screened out for body image disorder or other psychopathology. The Ss were diverse in occupation. For example, they were custodians, grounds keepers, building maintenance personnel, security guards, laboratory technicians, secretaries, administrators and professors. According to univariate analyses of variance, the four groups differed in age and education (see Table 2). In order of women and men, the statistics were: age, F = and ; education, F = and 47.59, P < See Table 2 for paired comparisons between groups. Within samples, the gender differences were: male Non-BDD Ss were slightly younger than female Non-BDD Ss, t = 1.97 (dr = 175), P = 0.05; both groups of male clinical Ss were lower in education than the women, t (df-- 80) and 3.66 (df = 175), P < 0.001; male university staff were higher in education than women, t = 3.57 (dr = 136), P < The groups also differed in marital status, Z 2 = (df = 6), P < The undergraduates were all single and a larger portion of the clinical Ss were single compared to the university staff. Regarding the racial make-up, the percentage of caucasian Ss was 97.7% for clinical Ss, 95.8% for undergraduates, and 94.4% for university staff. Some of the psychometric studies used subsamples of Ss. In such instances, the subsample was not significantly different from the remainder of the same sex Ss on the BDDE or any of the basic S characteristics. Measures Subjects read and signed a lay summary and consent form. The BDDE interview and questionnaires were administered by a BA-level clinical assistant. Body Dysmorphic Disorder Examination. The BDDE (version 3.2) as described in Study 1 was administered to all Ss. It was scored for the total score and the diagnosis of BDD. Higher scores represent more severe BDD symptoms. The interview questionnaire and rating forms are available from the authors. Multidimensional Body-Self Relations Questionnaire (MBSRQ). This body image measure was selected because it is sensitive to concerns about any aspect of appearance, not just weight or body shape, and has separate body image scales (Brown, Cash & Mikulka, 1990). Three out of 11 scales that are most relevant to attitudes about physical appearance were selected for analysis. (Scales pertaining to health and fitness attitudes, for example, were not analyzed.) The Appearance Evaluation subscale measures feelings of physical attractiveness or unattractiveness. The Body Areas Satisfaction Scale measures satisfaction or dissatisfaction with specific body areas. The Appearance Orientation Scale measures tendencies to groom and pay attention to one's appearance a great deal. The scales consist of 7, 8, and 12 items respectively, rated 1-5. Higher total scores for the first two scales represent more positive body image attitudes, whereas a lower score on Appearance Orientation is more positive. The rs for test-retest reliability on these scales range from 0.61 to 0.91 and from 0.85 to 0.88 for internal consistency. Concurrent validity coefficients with another measure of body satisfaction and the Appearance Evaluation Scale were 0.61 to 0.66.

6 760 James C. Rosen and Jeff Reiter Body Shape Questionnaire (BSQ). This is a 34-item questionnaire on body image (rated 1-6 for the past four weeks) that measures desire to lose weight, body dissatisfaction, thoughts of being too big or too fat, feelings of fatness after eating, and self-consciousness about weight in public (Cooper et al., 1987). The Body Shape Questionnaire was selected because it is one of the most widely used measures of negative body image in clinical studies. Because the measure is focused on stereotypic female weight and shape complaints, it was administered to women only. Higher scores represent more negative body image. The coefficients of internal consistency, test-retest reliability, and concurrent validity with other measures of body satisfaction are adequate: r = 0.97, 0.88, and 0.66, respectively (Rosen, Jones, Ramirez & Waxman, 1996). Brief Symptom Inventory. This measure was used in order to examine the relation between symptoms on the BDDE and psychopathology in general. We calculated the Global Severity Index which is the average severity of 53 psychological symptoms over the past week, rated 0-4. Test-retest reliability is r = 0.90 and internal consistency ranges from r = 0.71 to Validity coefficients with the MMPI are above 0.30 (Derogatis & Spencer, 1982). Rosenberg Self-Esteem Scale. This is a 10-item, 1-4 point scale that measures attitudes regarding general self-worth; higher ratings indicate more positive self-esteem (Rosenberg, 1979). The scale has acceptable test-retest reliability (0.77) and internal consistency (0.89), and it correlates significantly with peer ratings (r = 0.32; Demo, 1985). Internal consistency Procedure and results The Cronbach's alphas for the 28 items included in the total score were 0.81 (BDD Ss), 0.88 (Non-BDD clinical Ss), 0.93 (undergraduates), and 0.91 (university staff), indicating that the Body Dysmorphic Disorder Examination has good internal consistency. Test-retest reliability The Body Dysmorphic Disorder Examination was administered twice to 29 consecutive undergraduate Ss (one non-clinical S met diagnostic criteria for BDD) and 24 consecutive clinical Ss with a two-week interval between examinations. The same interviewer was used with each S. The test-retest reliability coefficients for the total score were, r = 0.94 and 0.87, P < 0.01, respectively. Thus, responses on the measure were stable over a short period of time. Correlations for each of the 28 symptom severity items were significant at P < 0.05 or less. Interrater reliability Of the non-clinical Ss, 25 consecutive university undergraduate Ss were interviewed and rated by one assistant, while a second, independent rater observed the interview. The reliability as measured by the intraclass correlation for total score on the BDDE was r = 0.99, P < 0.01, indicating good agreement between scores on the same interview, taking differences between raters into account. Interrater reliability also was tested on a sample of clinical Ss using two independent raters for separate interviews. The two interviews took place one to two weeks apart. Twenty consecutive clinical Ss (BDD and Non-BDD) were examined with the BDDE and asked to return for a reliability check. The intraclass correlation coefficient for the total score was r = 0.86, P < 0.01, indicating good agreement taking into account rater differences. Using the criteria on the Body Table 3. Correlations between the Body Dysmorphic Disorder Examination and measures of body image, psychological symptoms, and self-esteem Clinical Non-clinical Measure BDD Non-BDD University undergrads University staff Body Shape Questionnaire 0.60* 0.64* 0.77* 0.77* Appearance Evaluation * * * Body Areas Satisfaction * * * Appearance Orientation * 0.40* 0.44* Brief Symptom Inventory 0.44* 0.49* 0.37* 0.50* Rosenberg Self-Esteem * " * * *P < 0.05 after Bonferroni correction.

7 Body Dysmorphic Disorder Examination 761 Table 4. Logistic regressions predicting clinical or non-clinical status Predictors entered ~2 Probability Step I: Control variables Intercept < Age < Step 2: Psychological adjustment Intercept < Age < Self-esteem < Psychological symptoms Step 3: Body image questionnaires Intercept Age < Self-esteem < Psychological symptoms Body Areas Satisfaction Scale Appearance Evaluation Scale Appearance Orientation Scale Step 4: Body image questionnaires Intercept Age < Self-esteem I Psychological symptoms Body Areas Satisfaction Scale Appearance Evaluation Scale Appearance Orientation Scale Body Dysmorphic Disorder Exam < Dysmorphic Disorder Examination for BDD, 9 out of 20 Ss were diagnosed as BDD by the first examination. The agreement between the two examinations on BDD diagnosis, correcting for chance with Cohen's (Cohen, 1960) kappa was n = Concurrent validity The validity of the BDDE was evaluated with correlation coefficients between it and other measures of body image and psychological adjustment. A Bonferroni procedure was used to control for Type I error rate. There were no substantial gender differences in the correlations, so for convenience the results are reported by study sample with the male and female Ss combined (except for women only on the Body Shape Questionnaire). According to the correlations shown in Table 3, the BDDE total score was strongly associated with negative body image on the Body Shape Questionnaire, greater psychological symptoms and lower self-esteem in all four study samples. Compared to the Body Shape Questionnaire, the MBSRQ scales are more concerned with body dissatisfaction than preoccupation and distress. Those three body image scales had a weaker relation to BDD symptoms, especially with the BDD Ss (correlations with Body Satisfaction and Appearance Orientation were significant at P < 0.05 prior to Bonferroni correction). This suggests that at the more severe end of BDD symptoms, the BDDE reflects something other than strictly increased body dissatisfaction. The more restricted range of scores in the BDD group also reduced the magnitude of the correlations. Criterion group validity Clinical vs non-clinical status: univariate analysis. To test the criterion validity of the BDDE, we compared the total scores of the four groups (Table 2), controlling for age. According to the analyses of covariance, the groups differed significantly, F (3,262)= (males) and F (3,422) = (females), Ps < BDD Ss reported more severe symptoms on the BDDE than the three other groups. Non-BDD clinical Ss were more severe than the non-clinical controls and the undergraduate students reported more symptoms than the university staff. Regarding differences by gender, among BDD Ss, the men scored higher than the women (t = 2.94 (df = 80), P < 0.01), whereas the gender difference was reversed among the Non-BDD clinical Ss (t = 2.41 (df= 175), P = 0.018). The female students and staff scored higher on the BDDE than the corresponding male samples, t = 7.27 (df = 295) and 4.11 (df = 138), P < 0.001, respectively. Clinical vs non-clinical status: multivariate analysis. We then performed a hierarchical logistic regression analysis to test the incremental validity or independent contribution of the Body Dysmorphic Disorder Examination in predicting clinical (BDD and Non-BDD combined) or non-clinical (students and staff combined) group membership after controlling statistically for other

8 762 James C. Rosen and Jeff Reiter predictors. (Subjects were combined to form two groups in order to have an adequate number of Ss for this analysis.) We controlled first for any variance explained by age. Then, in the second step we entered psychological symptoms and self-esteem. This allowed us to determine if the body image measures added any information to discriminating the groups beyond that provided by general measures of psychopathology. Finally, we entered the MBSRQ scales together as a block in step three and the BDDE in step four. This tested the usefulness of the BDDE in classifying Ss after controlling for the type of body image information provided by the MBSRQ. Male and female Ss were combined. The Body Shape Questionnaire was not included as an independent variable because it was not administered to the male Ss. The results are shown in Table 4. The overall model at the first step provided a significant fit with the data, X 2 (1)= 45.81, P < , with age predicting S classification. Adding two measures of global psychological adjustment in the second step improved the classification, ;(2 (2) = , P < , though of the two, only the Rosenberg Self-Esteem Scale was a significant predictor of clinical status. In the third step, age and self-esteem remained significant predictors. Of the three MBSRQ body image scales that were added, the Appearance Orientation Scale was a significant predictor of clinical status. The improvement in S classification was significant, X 2 (3) = 28.13, P < In the fourth and final step, adding the BDDE led to an improvement in prediction, Z 2 (1) = 78.44, P < Age and self-esteem remained significant. None of the other body image scales contributed significantly to classifying Ss. The overall chi-square for the model was ;(2 (7) = , P < To double check the usefulness of the body image measures, we tested another model in which we reversed steps three and four above, so that the BDDE was entered in step three and the MBSRQ scales were entered in the final step. The predictors, at step three including the BDDE, were significant overall, ;(2 (4) = , P < At step four, the BDDE remained a significant predictor. Adding the other body image measures did not improve S classification, as the ~2 for improvement only gained 4.37 units, P = In conclusion, the information from the three other body image scales was redundant with the BDDE, which was better able to predict clinical status. Agreement with independent BDD diagnosis. We tested the validity of the BDDE in predicting a diagnosis of body dysmorphic disorder by comparing its classification of Ss as BDD using the criteria described above, with the diagnosis by the independent psychologist or psychiatrist. One-hundred and eighty-four clinic-referred Ss were evaluated by the diagnostician within two weeks of the administration of the BDDE (see Method section for clinical Ss above). The agreement between the two examinations on BDD diagnosis, correcting for chance with Cohen's (Cohen, 1960) kappa was = STUDY 3: TREATMENT VALIDITY The purpose of this study was to determine if the Body Dysmorphic Disorder Examination is able to detect changes resulting from interventions that are designed to reduce BDD symptoms. Method Subjects. The Ss were a subsample of the BDD referrals who participated in a controlled treatment trial of cognitive behavioral body image therapy (CBT) for BDD (the program and results are described in detail in Rosen, Reiter & Orosan, 1995a). They were 54 women with a mean age of 36.5 yr (SD = 9.3; range 21-61) and various appearance complaints (weight or body shape, facial features, breast size, skin blemishes, head hair, body hair, and teeth). Procedure. Subjects were randomly assigned to CBT or no-treatment control. Included in a battery of outcome measures was the BDDE, which the Ss completed at pre-treatment and post-treatment. The CBT Ss also were assessed at a four-month follow-up. Only the pre- to post-treatment results will be presented here, because the follow-up assessment was not experimentally controlled (control Ss were offered therapy after the post-treatment assessment). Patients were treated in groups for eight sessions. Therapy involved modification of intrusive thoughts of body dissatisfaction and overvalued beliefs about physical appearance, exposure to avoided body image situations and elimination of body checking. Results. Mean scores on the BDDE at pre- and post-treatment were 89.9 (SD = 16.1) and 83.2 (SD = 19.7) for the control Ss and 83.9 (SD = 14.8) and 41.1 (SD = 16.9) for the CBT Ss. Changes

9 Body Dysmorphic Disorder Examination 763 on the Body Dysmorphic Disorder Examination were analyzed in a between-group analysis of covariance at post-treatment, controlling for baseline scores on the BDDE. The group effect was significant, F (1,53) = 69.43, P < , indicating that BDDE scores were significantly lower in CBT Ss after treatment. We also calculated the rates of clinically significant improvement using the BDDE. To be considered clinically improved, a S had to: (a) no longer meet the diagnostic criteria on the BDDE and (b) have a score lower on the BDDE after treatment than their pre-treatment score minus two standard errors of measurement. Two Ss in the no-treatment condition (7.4%) met these criteria for improvement after the waiting period. Of the 27 cognitive therapy Ss, 22 were clinically improved at post-treatment (81.5%). Improvement on the BDDE was significantly correlated with improvement on the other outcome measures. Correlations (rs) between change scores (pre-treatment minus post-treatment) were: 0.69 for Body Shape Questionnaire, for Appearance Evaluation Scale, for Body Areas Satisfaction Scale, 0.49 for Appearance Orientation Scale, 0.59 for Brief Symptom Inventory, and for Rosenberg Self-Esteem Scale, all Ps < DISCUSSION The Body Dysmorphic Disorder Examination was shown in these studies to provide a reliable and valid measure of body dysmorphic disorder symptoms in BDD patients as well as in other clinical and non-clinical groups. More severe body dysmorphic symptoms on the BDDE were associated with more negative body image, more negative self-esteem, and more severe psychological symptoms. Though correlated with these variables, the BDDE possessed incremental validity because prediction of clinical status was more accurate when the BDDE was added to other body image measures. Divergent validity was not examined by testing the relation of the BDDE with measures theoretically unrelated to BDD symptoms. However, the BDDE did add unique information to the discrimination of clinical and non-clinical groups after taking into account more general measures of psychopathology and self-esteem. Used as an outcome measure, the BDDE showed reductions in body image symptoms after treatment of BDD and was sensitive to changes on other measures of body image and psychological adjustment. Two findings showed that the BDDE can differentiate BDD from other body image problems. First, the measure had good agreement with independent diagnoses of BDD in a clinical setting. Second, on average the BDD patients scored much higher on the measure than patients with other types of body image complaints and college women with body dissatisfaction. Although these studies represent an encouraging beginning to the assessment of body dysmorphic disorder, there are several issues that need further attention. We provided descriptive statistics on the BDDE; however, given limitations in the sample sizes and demographics, these statistics cannot be accepted as norms. In particular, it would be desirable to accumulate a larger sample of male Ss seeking treatment for BDD symptoms. Concerns about physical appearance are likely to vary by culture. Although at the present time there are no data to indicate that BDD in particular differs by ethnic identification, studies of the BDDE in larger samples of ethnic minorities would be appropriate. The BDDE criteria for the diagnosis of BDD seem adequate at present. However, the best operational definition of BDD should be revised as more information on the disorder is accumulated. The BDDE is time-consuming and requires a clinician to administer. This format might be problematic for clinicians or researchers who wish to measure body image disorder symptoms in large samples or in conjunction with a much more extensive battery of assessment procedures. Presently, we are testing a self-administration version of the BDDE that will address this limitation. However, based on our experience interviewing clinical patients, it is possible to administer the BDDE and a review of psychological symptoms and history in less than 1~ hr. Because the BDDE covers the typical BDD symptoms, we find that the interview helps the clinician to present him or herself as knowledgeable and empathetic and, consequently, the interview facilitates rapport and credibility. Moreover, to diagnose BDD, the diagnostician must use some form of direct examination in order to evaluate the person's physical appearance. On the other hand, a completely self-administered questionnaire would be acceptable for measuring strictly body image.

10 764 James C. Rosen and Jeff Reiter Although the BDDE was designed to distinguish BDD from other body image problems, it happens that the measure has good reliability and validity in persons with non-bdd body image complaints, and it correlates significantly with other dimensional measures of body image. Two other studies reported that the BDDE is reliable and valid with eating disorder Ss and weight-preoccupied women (Jackman, Williamson, Funsch & Warner, 1994; Rosen, Reiter & Orosan, 1995b). Therefore, the BDDE appears to be useful with different types of clinical samples and as both a diagnostic and dimensional measure. Having presented in this report some of the limitations of the BDDE, we believe the data are sufficient at this point to justify its use by other clinicians and investigators. We hope this measure will help to promote more systematic and objective study of body dysmorphic disorder. REFERENCES American Psychiatric Association (APA) (1987), Diagnostic and Statistical Manual of Mental Disorders--Revised (3rd edn). Washington, DC: APA. American Psychiatric Association (APA) (1994). Diagnostic and Statistical Manual of Mental Disorders (4th edn). Washington, DC: APA. Brown, T. A., Cash, T. F. & Mikulka, P. J. (1990). Attitudinal body-image assessment: factor analysis of the Body Self-Relations Questionnaire. Journal of Personality Assessment, 35, Chowdhury, A. N. (1989). Penile perception of Koro patients. Acta Psychiatrica Scandinavica 80, Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 10, Cooper, Z. & Fairburn, C. G. (1987). The Eating Disorder Examination: a semi-structured interview for the assessment of the specific psychopathology of eating disorders. International Journal of Eating Disorders, 6, 1-8. Cooper, P. J., Taylor, M. J., Cooper, Z. & Fairburn, C. G. (1987). The development and validation of the Body Shape Questionnaire. International Journal of Eating Disorders, 6, de Leon, J., Bott, A. and Simpson, G. M. (1989). Dysmorphophobia: body dysmorphic disorder of delusional disorder, somatic subtype? Comprehensive Psychiatry 30, Demo, D. H. (1985). The measurement of self-esteem: refining our methods. Journal of Personality and Social Psychology, 48, Derogatis, L. R. & Spencer, P. M. (1982). The Brief Symptom Inventory: Administration, Scoring, and Procedures Manual. Baltimore: Clinical Psychometric Research. Franzoi, S. L. & Shields, S. A. (1984). The Body Esteem Scale: multidimensional structure and sex differences in a college population. Journal of Personality Assessment, 48, Garner, D. M. (1991). Eating Disorder Inventory--2. Odessa, FL: Psychological Assessment Resources. Hollander, E., Cohen, L. J. & Simeon, D. (1993). Body dysmorphic disorder. Psychiatric Annals, 23, Jackman, L. P., Williamson, D. A., Funsch, C. L. & Warner, M. S. (1994). Body dysmorphic disorder in female college athletes. Manuscript submitted for publication. Jerome, L. (1992). Body dysmorphic disorder: a controlled study of patients requesting cosmetic rhinoplasty (letter). American Journal of Psychiatry, 149, 577. Morselli, E. (1886). Sulla Dismorfofobia e sulla tafefobia. Bolletino Accademia delll Scienze Mediche di Genova, 6, Neziroglu, F. A. & Yaryura-Tobias, J. A. (1993a). Exposure, response prevention, and cognitive therapy in the treatment of body dysmorphic disorder. Behavior Therapy, 24, Neziroglu, F. A. & Yaryura-Tobias, J. A. (1993b). Body dysmorphic disorder: phenomenology and case descriptions. Behavioural Psychotherapy, 21, Phillips, K. A. (1991). Body dysmorphic disorder: the distress of imagined ugliness. American Journal of Psychiatry, 148, i Phillips, K. A. (1993). Body Dysmorphic Disorder Modification of the YBOCS, McLean Version. Belmont, MA: McLean Hospital. Phillips, K. A. (1994). Body Dysmorphic Disorder Diagnostic Module. Belmont, MA: McLean Hospital. Phillips, K. A. & McElroy, S. L. (1993). Insight, overvalued ideation, and delusional thinking in body dysmorphic disorder: theoretical and treatment implications. Journal of Nervous and Mental Disease, 181, Phillips, K. A., McElroy, S. L., Keck, P. E., Pope, H. G. & Hudson, J. I. (1993). Body dysmorphic disorder: 30 cases of imagined ugliness. American Journal of Psychiatry, 150, Rosen, J. C. (1995). The nature of body dysmorphic disorder and treatment with cognitive behavior therapy. Cognitive and Behavioral Practice, 2, Rosen, J. C., Jones, A., Ramirez, E. & Waxman, S. (1996). Body Shape Questionnaire: studies of validity and reliability. International Journal of Eating Disorders, 20, in press. Rosen, J. C., Reiter, J. & Orosan, P. (1995a). Cognitive behavioral body image therapy for body dysmorphic disorder. Journal of Consulting and Clinical Psychology, 63, Rosen, J. C., Reiter, J. & Orosan, P. (1995b). Assessment of body image in eating disorders with the Body Dysmorphic Disorder Examination. Behaviour Research and Therapy, 33, Rosenberg, M. (1979). Conceiving the Self. New York: Basic Books. Williamson, D. A., Davis, C. J., Bennett, S. M., Gorenczny, A. J. & Gleaves, D. H. (1989). Development of a simple procedure for assessing body image disturbances. Behavioral Assessment, 11, APPENDIX (See opposite)

11 Body Dysmorphic Disorder Examination 765.-~ ~ ~.~ o ~ -~,~,~.~.~ "o.=.-.~ ~.~ -I - ~ ~ ~ ; =,~ ~.- ~. ~.o.-.. o. o o E 6 % -~o ~.-~ ~- ~ ~-~ ~ ~ -..~ >', ~ ~ ~ II > II ~.. ~". ~ --,,~ ~ - ~ ~ - - ~ ~-.= O I~ ~ O I:: ~ ~ ~.~ ~ - -..~ ~. UJ r~ r~ :~'~'~ ~; ~"~.e ~' "~ ~"~;~ ~"~ ~'~ ~ O ~ O ~. ~ O ~ ~.- = = i..= ~=.~.~ ~ ~..~.~ ~:..= ~ ~ ~ 3 ~ ' ~ II II,-~ il II,.~ li II II ~ II '- II.~ II [I "~,. ii II II II II ~,~.~ ~. ~. ~ ~= ~ - ~ ~ :

12 766 James C. Rosen and.left Reiter o o o ~" ~ ~ ~.,-,o ~ ~ -~,, ~.~,, ~ ~g