Somatoform dissociation in eating-disordered patients

Behaviour Research and Therapy 41 (2003) 619–627 www.elsevier.com/locate/brat

Somatoform dissociation in eating-disordered patients G. Waller a,∗, M. Babbs a, F. Wright a, C. Potterton a, C. Meyer b, N. Leung c a

c

Department of Psychiatry, St. George’s Hospital Medical School, University of London, Cranmer Terrace, London SW17 0RE, UK b Department of Psychology, University of Warwick, Coventry CV8 1NE, UK Eating Disorders Service, South Birmingham Mental Health NHS Trust, Queen Elizabeth Hospital, Edgbaston, Birmingham, UK Received 10 July 2002; received in revised form 9 December 2002; accepted 11 December 2002

Abstract This study investigated the role of somatoform dissociation in eating disorders and pathological eating behaviour, relative to the established association of eating pathology with psychological dissociation. The participants were 131 women with DSM-IV diagnoses of anorexic or bulimic disorders and 75 women who had no such disorder. Each woman completed measures of psychological and somatoform dissociation, as well as a measure of bulimic attitudes. The current presence or absence of specific bulimic behaviours was identified during the clinical interview. Levels of both forms of dissociation were higher in the women who had diagnoses of disorders with a bulimic component (bulimia nervosa; anorexia nervosa of the binge/purge subtype) than in the non-clinical or restrictive anorexic women. Somatoform dissociation showed particularly strong links with the presence of bulimic behavioural features (excessive exercise, laxative abuse, diet pill abuse, diuretic abuse) and with bulimic attitudes. The formulation and treatment of cases where there are bulimic features is likely to be enhanced by the assessment of somatoform dissociation.  2003 Elsevier Science Ltd. All rights reserved. Keywords: Eating disorder; Dissociation; Bulimic behaviours; Somatoform

1. Dissociation in the eating disorders Dissociation has been described as involving a failure to process and integrate information that would normally be perceived and conceived in an integrated way (Spiegel & Carden˜a, 1991). It

∗

Corresponding author. Tel.: +44-208-725-5543; fax: +44-208-725-2618. E-mail address: [email protected] (G. Waller).

0005-7967/03/$ - see front matter  2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0005-7967(03)00019-6

620

G. Waller et al. / Behaviour Research and Therapy 41 (2003) 619–627

is commonly reported in cases of some Axis I and Axis II psychopathologies (Carlson & Putnam, 1993), including the eating disorders (e.g., Everill, Waller, & Macdonald, 1995; Vanderlinden, Vandereycken, van Dyck & Vertommen, 1993). This pattern of cognitive and emotional avoidance in bulimia can be seen as similar to the ‘escape from awareness’ model of eating behaviour (Heatherton & Baumeister, 1991), where overeating is a product of the disinhibition caused by avoidance of threat processing. It might also be seen as related to the pattern of overeating to block awareness of intolerable emotional states, particularly following trauma (Root & Fallon, 1989). Such dissociation is particularly prominent among individuals who report a history of trauma in childhood or adulthood (e.g., Chu & Dill, 1990; Lipschitz, Kaplan, Sorkenn, Chorney, & Asnis, 1996), acting as a defensive process for escaping awareness of intolerable experiences that cannot be avoided through fight or flight. Until recently, dissociation has been conceptualised and measured largely as a cognitive/emotional disruption, manifesting in experiences such as absorption, amnesia, depersonalization and derealization (Carlson & Putnam, 1993). However, Nijenhuis, Spinhoven, Vanderlinden, van Dyck, & van der Hart (1998) have suggested that this formulation of a cognitive– emotional construct (psychological dissociation) is insufficient, and that there is a need to expand that model to include a physiological component (somatoform dissociation). There is a range of evidence that repeated and inescapable trauma has such physiological correlates, particularly neurophysiological reactivity and blunted autonomic nervous system responses to threat and harm (e.g., Carrey, Butter, Persinger, & Bialek, 1995; Griffin, Resick, & Mechanic; 1997; van der Kolk et al., 1996). While psychological dissociation is linked with a wide range of traumatic experiences (including emotional abuse and neglect), somatoform dissociation is specifically associated with physically-based forms of trauma, such as physical abuse and contact sexual abuse (Nijenhuis, Spinhoven, van Dyck, van der Hart & Vanderlinden, 1998; Waller et al., 2000). It includes the experience of physiological features such as anaesthesia, analgesia and behavioural freezing. Nijenhuis, Spinhoven, van Dyck, van der Hart and Vanderlinden (1996) have developed a wellvalidated self-report measure of these physiological characteristics—the Somatoform Dissociation Questionnaire (SDQ-20). There are phenomenological links between the eating disorders and different forms of trauma, including contact and non-contact forms (e.g., Fallon & Wonderlich, 1997; Kent & Waller, 2000). Given the diverse nature of those links, it is not possible to conclude whether dissociation in eating disorders is best understood in terms of the psychological or the physiological variant, or whether both are relevant. To date, the literature on dissociation in eating disorders has focused on the psychological form, and there is no evidence regarding the role of somatoform dissociation across the range of eating disorders. The aim of this study is to determine the links between different forms of dissociation and eating psychopathology. First, it is hypothesised that both forms of dissociation will be stronger in those women with a bulimic disorder than among restrictive anorexics. Second, it is anticipated that both forms will be stronger among eating-disordered patients who experience bulimic symptoms (bingeing, purging and exercise). Finally, it is hypothesised that the two forms of dissociation will be positively associated with a dimensional measure of bulimic pathology.

G. Waller et al. / Behaviour Research and Therapy 41 (2003) 619–627

621

2. Method 2.1. Design A comparative design is used to examine levels of dissociation among eating-disordered patients and in non-clinical women. Comparisons were made across diagnoses, and according to the presence or absence of specific bulimic behaviours. Regression analysis was used to determine the dimensional links of bulimic pathology with different forms of dissociation. 2.2. Participants The clinical group consisted of 131 women who met strict DSM-IV (American Psychiatric Association, 1994) diagnostic criteria for eating disorder diagnoses. Diagnosis was made using a structured interview, based on DSM-IV criteria (see below). Weight and height were measured at assessment, and were used to calculate body mass index (BMI = weight[kg] / height[m]2). Those diagnoses were: anorexia nervosa of the restrictive subtype (N = 21; BMI = 16.7, SD = 2.95), anorexia nervosa of the binge/purge subtype (N = 40; BMI = 17.0, SD = 2.03); and bulimia nervosa (N = 70; BMI = 24.4, SD = 8.57). A one-way ANOVA (with post hoc Tukey’s tests) showed the expected difference in BMI (F = 21.6; df = 2,128; P ⬍ 0.001) between the two anorexic groups and the bulimia nervosa patients (P ⬍ 0.05 in each case), but no difference between the anorexic groups. These women were referrals to two specialist tertiary level eating disorders services, which served both local and national catchments. A further 29 referrals were excluded from this study because they met criteria for an Eating Disorder Not Otherwise Specified (EDNOS, including binge eating disorder) or because they did not meet eating disorder diagnostic criteria. No diagnosable males were referred in the course of the study. There were no data available on comorbid states, including mood disorders, impulsive disorders, and personality disorders. The non-clinical group consisted of 75 women, who were interviewed by the researchers to exclude any DSM-IV diagnosis of an eating disorder (including EDNOS), using the structured interview reported above. In addition, all participants completed the Bulimic Investigatory Test, Edinburgh (BITE; see below). The suggested cut-off score on this measure is 25 (Henderson & Freeman, 1987), indicating clinical caseness in a range of eating disorders (rather than just in bulimia nervosa—Waller, 1992). In order to reduce the risk of including unidentified cases, any non-clinical individual who scored above 15 on the overall BITE scale was excluded. Using the criteria outlined above, a further 14 participants were excluded from this group. No objective BMI data were available for this group. Table 1 shows that the clinical groups were well matched for age, but that they were significantly older than the non-clinical women. However, age was not associated with BITE, DES-II or SDQ-20 scores in either the clinical or the non-clinical group (P ⬎ 0.10 in all cases), with the exception of the DES-II scores in the clinical group (r (df = 131) = ⫺0.196, P ⬍ 0.03). Therefore, age is not reported as a covariate in the following analyses (although its inclusion made no difference to the findings).

a

P⬍0.001.

Age (years) Bulimic Investigatory Test [BITE] Cognitive dissociation [DESII] Somatoform dissociation [SDQ20]

N

1.31 (0.35)

1.35

16.9

13.4

(9.50)

21 25.7 17.7

(SD)

(0.38)

(15.1)

(9.06) (10.2)

Anorexia nervosa (restrictive) M (SD)

75 20.1 (1.28) 8.87 (8.00)

M

Group Non-clinical

1.91

26.7

40 28.0 30.3

(0.74)

(19.8)

(6.83) (10.6)

Anorexia nervosa (binge-purge) M (SD)

1.63

20.3

70 26.9 36.7

M

(0.58)

(16.2)

(6.56) (7.77)

(SD)

Bulimia nervosa

11.9a

7.35a

24.1a 134.4a

NC⬍BN⬍BA; RA⬍BA

NC⬍BA=BN

NC⬍RA=BA=BN NC⬍RA⬍BA⬍BN

ANOVA F Tukey tests (P ⬍0.05)

Table 1 Age, bulimic psychopathology and levels of dissociation of non-clinical women (NC) and female patients with restrictive anorexia nervosa (RA), anorexia nervosa of the binge/purge subtype (BA) and bulimia nervosa (BN)

622 G. Waller et al. / Behaviour Research and Therapy 41 (2003) 619–627

G. Waller et al. / Behaviour Research and Therapy 41 (2003) 619–627

623

2.3. Measures and procedure The non-clinical participants were volunteers, who received no payment for taking part. They were recruited from a mixed group of undergraduates and non-students. The patients (who also received no payment for participating) were all assessed for specialist in- or out-patient treatment for eating disorders. Diagnoses were made by experienced clinical psychologists and psychiatrists, who were trained in the application of DSM-IV diagnoses. The diagnostic process included questions about whether the patients currently (over the last three months and within the last two weeks) engaged in the following eating and weight control behaviours: objective bingeing (including a large amount of food and a sense of loss of control); vomiting for weight loss; abuse of laxatives; abuse of diuretics; abuse of diet pills; or excessive exercise (a minimum of six hours exercise per week, with the intention of weight control or reduction). In cases where there was any uncertainty about diagnosis, the matter was resolved through team discussion of the case (sometimes involving further assessment by another clinician). As part of their assessment, the patients completed a number of standardized psychometric questionnaires to evaluate their psychopathology and treatment outcome. These measures were completed in the order given below. All diagnoses were made without knowledge of the outcome of the psychometric measures. The Dissociative Experiences Scale—Revised version (DES-II; Carlson & Putnam, 1993) is a 28-item self-report scale, where higher scores reflect more severe psychological dissociation (including experiences of amnesia, depersonalization/derealization and absorption). Each item is rated on a 0–100% scale, and the individual’s score (range = 0–100) is calculated by taking the mean score of the 28 items. The DES-II has good overall psychometric properties, but factor analysis shows that it lacks a consistent sub-scale structure (Carlson & Putnam, 1993). Therefore, the overall score was used. The Somatoform Dissociation Questionnaire (SDQ-20) was devised by Nijenhuis, Spinhoven, van Dyck, van der Hart and Vanderlinden (1996). Higher item-mean scores on the SDQ-20 ( range = 1–5) reflect greater levels of somatoform dissociation. This construct involves the presence and severity of physiological features (e.g., analgesia, pain) that have been hypothesised to be ‘freeze’ responses to imminent and inescapable physical threat (Nijenhuis et al., 1998). Sample items include: “It is as if my body, or part of it, has disappeared”; “My body, or part of it, is insensitive to pain”; “My body, or a part of it, feels numb”; and “I do not have a cold but yet am able to smell much better or worse than I usually do”. Each of the 20 items is rated on a five-point Likert scale. Nijenhuis and colleagues (Nijenhuis et al., 1996; Nijenhuis et al., 1998; Nijenhuis et al., 1999) have shown that the scale has acceptable clinical and psychometric characteristics, with good internal consistency, concurrent validity and convergent validity. Given that this measure has not been widely used with eating-disordered patients, the psychometric properties of the scale were considered here. Inspection of the scree plot from a principal components analysis (unrotated) showed that the patients’ SDQ-20 responses had a single-factor solution (eigenvalue = 7.19, explained variance = 36.0%), where all items loaded onto that factor (coefficients = 0.438–0.774). Internal consistency of the scale was high (Cronbach’s alpha = 0.899). There were not sufficient participants in the non-clinical group to undertake a similar factor analysis, but the internal consistency score on the whole SDQ-20 for this group was similar (Cronbach’s alpha = 0.825). These psychometric characteristics are comparable with those reported previously (Nijenhuis et al., 1996), and support the use of the SDQ-20 in its current form.

624

G. Waller et al. / Behaviour Research and Therapy 41 (2003) 619–627

Higher scores on the Bulimic Investigatory Test—Edinburgh (BITE; Henderson & Freeman, 1987) reflect stronger eating-disordered attitudes and behaviours. Despite its ostensible purpose (detection of bulimic pathology), the BITE was used with these groups because it is a good measure for discriminating eating disorder diagnostic groups (including restrictive anorexics) from each other and from non-clinical women (Waller, 1992). Other dimensional measures of general eating pathology tend to be poorer at differentiating diagnostic groups (Gross, Rosen, Leitenberg, & Willmuth, 1986). The total BITE score is reported in this study, since it reflected the findings from the two BITE subscales (Symptom and Severity). 2.4. Data analysis The data were sufficiently normally distributed to justify the use of parametric analyses. First, the groups’ levels of eating pathology (BITE) and dissociative characteristics (DES-II, SDQ-20) were compared using one-way ANOVAs, with post-hoc Tukey’s tests (P ⬍ 0.05) to determine pairwise differences. Second, among the 110 patients with either anorexia nervosa of the bingepurge subtype or bulimia nervosa, the levels of dissociation (DES-II, SDQ-20) in those who did or did not currently use individual bulimic behaviours (bingeing, vomiting, excessive exercise, and abuse of laxatives, diuretics and diet pills) were compared using t-tests. Finally, the dimensional associations between dissociation (DES-II, SDQ-20) and bulimic pathology (BITE) were tested using correlations (Pearson’s r) and multiple regression analyses. 3. Results 3.1. Levels of dissociation in non-clinical and eating-disordered groups Table 1 shows the women’s mean ages, and their scores on the measures of bulimic psychopathology (BITE), cognitive dissociation (DES-II) and somatoform dissociation (SDQ-20). The groups were well matched for age, and showed the expected differences in BITE scores (Waller, 1992). The restrictive anorexics did not differ from the non-clinical women on either dissociation measure, but both bulimic groups had higher levels of both forms of dissociation than the non-clinical women. The bulimic anorexics had particularly high levels of somatoform dissociation, scoring higher than both of the other clinical groups. 3.2. Categorical links between dissociation and specific bulimic symptoms There were 110 women (i.e., 70 bulimia nervosa and 40 anorexia nervosa of the binge-purge subtype) in the sample who could be described as ‘bulimic’ (i.e., having disorders that involved bingeing and/or purging behaviour). The behaviours used included: objective bingeing (N = 89); vomiting (N = 86); exercising excessively for weight loss (N = 52); and abusing laxatives (N = 42), diet pills (N = 10) and diuretics (N = 8). Levels of dissociation were compared across those who did and those who did not report each of these behaviours, using independent samples t-tests. The presence of bingeing and vomiting behaviours were not associated with a higher level of dissociation, using either scale (t ⬍ 1.40, NS in all cases). In addition, correlational analyses

G. Waller et al. / Behaviour Research and Therapy 41 (2003) 619–627

625

(Pearson’s r) showed that the women’s frequencies of binging and vomiting were not reliably associated with either psychological dissociation (r = 0.08 and 0.09, respectively) or somatoform dissociation (r = 0.09 and 0.07, respectively). Somatoform dissociation was significantly higher among those who abused laxatives (M = 1.93, SD = 0.77 vs M = 1.60, SD = 0.56; t = 2.21, P ⬍ 0.04), diuretics (M = 2.34, SD = 1.34 vs M = 1.69, SD = 0.60; t = 2.19, P ⬍ 0.04) and diet pills (M = 2.46, SD = 1.35 vs M = 1.68, SD = 0.59; t = 2.64, P ⬍ 0.01). In contrast, psychological dissociation was associated with none of those three behaviours (t ⬍ 1.85, NS in all cases). Finally, excessive exercise was associated with higher levels of both psychological (M = 27.7, SD = 18.2 vs M = 18.8, SD = 16.9; t = 2.55, P ⬍ 0.02) and somatoform dissociation (M = 1.97, SD = 0.71 vs M = 1.53, SD = 0.55; t = 3.43, P ⬍ 0.001). 3.3. Dimensional links between dissociation and eating attitudes Dimensional associations were examined between different forms of dissociation (DES-II and SDQ-20) and eating attitudes (BITE). Since scores on the two measures of dissociation were positively correlated in each of the four groups (control group - r = 0.53, P ⬍ 0.001; restrictive anorexia nervosa - r = 0.69, P ⬍ 0.001; binge/purge anorexia nervosa - r = 0.77, P ⬍ 0.001; bulimia nervosa - r = 0.75, P ⬍ 0.001), multiple regression analyses were used to determine the most parsimonious model of the dissociation-eating link in each group. Among the non- clinical women, there was a significant overall link (F = 5.28, P ⬍ 0.01, explained variance = 10.4%), which was a product of a significant association with somatoform dissociation (t = 3.02, P ⬍ 0.01, beta = 0.39) but not psychological dissociation (t = 0.58, NS, beta = ⫺0.08). A similar pattern was found among the non-clinical group, where there was a significant overall link (F = 3.56, P ⬍ 0.05, explained variance = 4.1%). Again, this was a product of a significant association with somatoform dissociation (t = 2.07, P ⬍ 0.05, beta = 0.29) but not psychological dissociation (t = 0.56, NS, beta = ⫺0.07). Therefore, understanding the dimensional relationship between dissociation and eating pathology appears to depend on the role of somatoform dissociation. 4. Discussion This study has examined levels of psychological and somatoform dissociation across the eating disorders. At the diagnostic level, restrictive anorexics had relatively normal levels of each type of dissociation, while the binge–purge anorexics had the highest levels. Considering individual behaviours, the presence of specific purging and compensatory behaviours (exercise and the abuse of laxatives, diuretics and diet pills) was linked to higher levels of somatoform dissociation, although some were only infrequently reported as current behaviours. Somatoform dissociation was also the better predictor of eating attitudes in the eating-disordered and the non-clinical groups. Therefore, somatoform dissociation appears to be the better distinguishing characteristic of purging behaviours in the eating disorders. It is noteworthy that the relevant behaviours are those characterised by recent research (Vaz et al., 2001; Waller et al., 2001) as being ‘slowacting’ and as being linked to emotional avoidance. Therefore, it is possible that the physiological element of dissociation (marked by reduced autonomic responsivity, possibly consequential on

626

G. Waller et al. / Behaviour Research and Therapy 41 (2003) 619–627

trauma—e.g., Griffin, Resick & Mechanic, 1997) is a key factor in the use of strategies that have a long-term emotional inhibition function. Alternatively, the reduced autonomic activity, the dissociation and the purging behaviours might form a cluster of physiological, cognitive and behavioural defense mechanisms against emotional overarousal. These findings are limited by the lack of comorbidity data. There is a clear need to determine the relevance of each form of dissociation to different psychiatric and behavioural disorders, and to determine whether the relationship with eating pathology might be better explained as a covariate of other impulsive and affective problems. The findings also have the limitation of being based on the use of self-report of bulimic behaviours at interview. Such self-report has been suggested to be overinflated relative to diary records (e.g., Fahy & Eisler, 1993), particularly among patients with high levels of impulsivity (who are also likely to experience high levels of dissociation). Future research should use contemporaneous records of bulimic behaviours, to reduce any such effect. It should also test the hypothesis that somatoform dissociation in this group will be associated specifically with a history of physically-based traumas, rather than emotional abuse or non-contact trauma (Nijenhuis et al., 1998; Waller et al., 2000). Such developments will depend on developing a fuller understanding of the physiology-behaviour links in somatoform dissociation (Carrey et al., 1995; Griffin et al., 1997; van der Kolk et al., 1996), drawing from the literatures on learned helplessness and biological freezing (Nijenhuis et al., 1998). These findings suggest that the treatment of the eating disorders might need to address somatoform as well as (or instead of) psychological dissociation, especially where the disorder involves ‘slow-acting’ compensatory behaviours (e.g., laxative abuse). Methods that have been suggested for the treatment of dissociation (e.g., Kennerley, 1996; Mollon, 1996) might be of limited effectiveness unless they are extended to address the physiological aspect. This might include pharmacotherapy or prolonged exposure and habituation to affective cues (e.g., Waller et al., 2000). While somatoform dissociation includes a number of physiological-behavioural clusters (anaesthesia– analgesia, urogenital pain, freezing), it appears that the analgesic cluster is the central element (Nijenhuis, Spinhoven, Vanderlinden et al., 1998). Thus, it will be important to develop methods to reduce this part of the dissociative experience, both in the eating disorders and in other psychiatric disorders. References American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington: American Psychiatric Association. Carlson, E. B., & Putnam, F. W. (1993). An update on the Dissociative Experiences Scale. Dissociation, 6, 16–27. Carrey, N. J., Butter, H. J., Persinger, M. A., & Bialek, R. J. (1995). Physiological and cognitive correlates of child abuse. Journal of the Academy of Child and Adolescent Psychiatry, 34, 1067–1075. Chu, J. E., & Dill, D. L. (1990). Dissociative symptoms in relation to childhood physical and sexual abuse. American Journal of Psychiatry, 147, 887–892. Everill, J., Waller, G., & Macdonald, W. (1995). Dissociation in bulimic and non-eating-disordered women. International Journal of Eating Disorders, 17, 127–134. Fahy, T., & Eisler, I. (1993). Impulsivity and eating disorders. British Journal of Psychiatry, 162, 193–197. Fallon, P., & Wonderlich, S. A. (1997). Sexual abuse and other forms of trauma. In D. M. Garner, & P. E. Garfinkel (Eds.), Handbook of treatment for eating disorders. 2nd ed. (pp. 394–414). New York: Guilford.

G. Waller et al. / Behaviour Research and Therapy 41 (2003) 619–627

627

Griffin, M. G., Resick, P. A., & Mechanic, M. B. (1997). Objective assessment of peritraumatic dissociation: Psychophysiological indicators. American Journal of Psychiatry, 154, 1081–1088. Gross, J., Rosen, J. C., Leitenberg, H., & Willmuth, M. E. (1986). Validity of the Eating Attitudes Test and the Eating Disorders Inventory in bulimia nervosa. Journal of Consulting and Clinical Psychology, 54, 875–876. Heatherton, T. F., & Baumeister, R. F. (1991). Binge eating as an escape from self- awareness. Psychological Bulletin, 110, 86–108. Henderson, M., & Freeman, C. P. (1987). A self-rating scale for bulimia. The ‘BITE’. British Journal of Psychiatry, 150, 18–24. Kennerley, H. (1996). Cognitive therapy of dissociative symptoms associated with trauma. British Journal of Clinical Psychology, 35, 325–340. Kent, A., & Waller, G. (2000). Childhood emotional abuse and eating psychopathology. Clinical Psychology Review, 20, 887–903. Lipschitz, D. S., Kaplan, M. L., Sorkenn, J., Chorney, P., & Asnis, G. M. (1996). Childhood abuse, adult assault, and dissociation. Comprehensive Psychiatry, 37, 261–266. Mollon, P. (1996). Multiple selves, multiple voices: Working with trauma, violation and dissociation. Chichester, UK: Wiley. Nijenhuis, E. R. S., Spinhoven, P., Vanderlinden, J., van Dyck, R., & van der Hart, O. (1998). Somatoform dissociative symptoms as related to animal defensive reactions to predatory imminence and injury. Journal of Abnormal Psychology, 107, 63–73. Nijenhuis, E. R. S., Spinhoven, P., van Dyck, R., van der Hart, O., & Vanderlinden, J. (1996). The development and psychometric characteristics of the Somatoform Dissociation Questionnaire (SDQ-20). Journal of Nervous and Mental Disease, 184, 688–694. Nijenhuis, E. R. S., Spinhoven, P., van Dyck, R., van der Hart, O., & Vanderlinden, J. (1998). Degree of somatoform and psychological dissociation in dissociative disorder is correlated with reported trauma. Journal of Traumatic Stress, 11, 711–730. Nijenhuis, E. R. S., van Dyck, R., Spinhoven, P., van der Hart, O., Chatrou, M., Vanderlinden, J., & Moene, F. (1999). Somatoform dissociation discriminates among diagnostic categories over and above general psychopathology. Australian and New Zealand Journal of Psychiatry, 33, 511–520. Root, M. P. P., & Fallon, P. (1989). Treating the victimized bulimic. Journal of Interpersonal Violence, 4, 90–100. Spiegel, D., & Carden˜ a, E. (1991). Disintegrated experience: The dissociative disorders revisited. Journal of Abnormal Psychology, 100, 366–378. Van der Kolk, B. A., Pelcovitz, D., Roth, S., Mandel, F. S., McFarlane, A., & Herman, J. L. (1996). Dissociation, somatization, and affect dysregulation: The complexity of adaptation to trauma. American Journal of Psychiatry, 153 (Festschrift suppl), 83–93. Vanderlinden, J., Vandereycken, W., van Dyck, R., & Vertommen, H. (1993). Dissociative experiences and trauma in eating disorders. International Journal of Eating Disorders, 13, 187–193. Vaz, F. J., Pen˜ as, E. M., Ramos, M. I., Lopez-Ibor, J. J., & Guisado, J. A. (2001). Subtype criteria for bulimia nervosa: Short- versus long-term compensatory behaviours. Eating Disorders, 9, 301–311. Waller, G. (1992). Bulimic eating attitudes in different eating disorders. International Journal of Eating Disorders, 11, 73–78. Waller, G., Hamilton, K., Elliott, P., Lewendon, J., Stopa, L., Waters, A., Kennedy, F., Lee, G., Pearson, D., Kennerley, H., Hargreaves, I., Bashford, V., & Chalkley, J. (2000). Somatoform dissociation, psychological dissociation and specific forms of trauma. Journal of Trauma and Dissociation, 1, 81–98. Waller, G., Meyer, C., & Milligan, R. (2001). The importance of purging in the eating disorders: Time for a rethink. Paper presented at the Eating Disorders Research Society, Bernalillo NM, November.