• Return to Medical Papers

    FULL PAPER (日本語ー抄録のみ)


    D. Berger, M.D. (1,2), Y. Ono, M.D. (3),S. Saito, M.D.(2), I. Tezuka, B.A.(2), Y. Takahashi, M.D. (2), M. Uno, M.D. (2), Y. Ishikawa, M.D.(2), M. Asai, M.D.(3), H. Suematsu, M.D.(4), T. Kuboki, M.D. (4)

    Running title: Parental bonding, dissociation, and child abuse in Japanese eating disorders

    Key Words: Parental bonding, dissociation, child abuse, eating disorders, Japan

    Acta Psychiatrica Scand, 91:278-282; 1995.

    PDF Format:

    1, Albert Einstein College of Medicine, Bronx, New York, U.S.A.; 2, Tokyo Institute of Psychiatry, Tokyo, Japan; 3, Keio University Department of Psychiatry, Tokyo, Japan; 4, Tokyo University Department of Psychosomatic Medicine, Tokyo, Japan,

    Address reprint requests to Dr. Douglas Berger, Department of Psychopathology, Tokyo Institute of Psychiatry, 2-1-8, Kamikitazawa, Setagayaku, Tokyo Japan 156.

    Presented in part at the Japanese Stress Science Conference, Tokyo, Japan, December 1992. Supported through the National Institutes of Health Foreign Funded Research Program, Bethesda, Maryland U.S.A., by the Japanese Society For The Promotion of Science, Tokyo Japan.


    Parental bonding patterns were studied in 52 female Japanese eating disorder outpatients with and without histories of sexual or physical abuse, and with dissociation. Instruments included the Parental Bonding Instrument (PBI), the Dissociative Experiences Scale (DES), and the Dissociative Disorders Interview Schedule (DDIS). Those with physical abuse history, but not sexual abuse history, had significantly different parental bonding scores and higher DES scores compared to subjects without physical abuse. DES scores and PBI scores were not corelated. While the PBI was useful in discriminating between those with and those without abuse histories, it did not detect differences in degree of dissociation. Lack of association of sexual abuse to PBI and DES scores may have been due to mild abuse.


    While a handful of studies (1-7) have reported on the relationship of parental bonding patterns in eating disorders as measured by the Parental Bonding Instrument (PBI) (8), none have looked for differences between those with and without histories of child abuse. A number of studies have shown that eating disorders have an association with childhood abuse (9-12), and it seems likely that a history of childhood abuse would be reflected by the pattern of bonding with parents. The PBI may be a suitable instrument for this because it specifically focuses on two axes of bonding: care and protection.

    The only study that mentioned childhood abuse in their use of the PBI looked at PBI scores in an inpatient borderline group (13), 86.7% who had histories of sexual abuse. This study found significantly lower care (especially low paternal care) and more overprotection from both parents when compared to normal population data. To date, no study has compared the PBI scores of those with abuse histories to those without in any diagnostic sample.

    Studies that have looked at childhood abuse histories and dissociation in eating disorders have found higher levels of dissociation in the group with abuse histories (14, 31) and there has been a significant literature on the association of eating disorders with dissociation (15-22). These studies suggest that trauma-induced dissociation may be a significant factor in symptom development in a subgroup of eating disorders. Up to now, no studies have looked at the relationship of PBI scores with dissociation.

    The anecdotal stereotype of the overbearing mother and the ineffectual father in anorexia (1), as well as the formulation of the families of eating disordered as being rigid, overinmeshed and overprotective (23) have been studied with the PBI (1-7). Results of these studies have found varying patterns of significant differences on the care and protection scales for each parent when compared to controls (2-7), though notably some (2,3,6,7) found evidence of paternal overprotectiveness tending to go against the picture of the ineffectual father.

    The present study sought to investigate the relationship of parental bonding patterns as measured by the PBI with abuse histories and dissociation, and was a pilot attempt to determine the usefulness of the PBI in the assessment of parental bonding of those with abuse histories and/or dissociation. Considering the high rate of abuse and dissociation in eating disorders, failure to take these variables into account in studies of parental bonding in eating disorders may be one cause for some of the variability in the results. Characterizing the relationship patients with abuse histories and/or significant dissociative symptoms have with their parents may help with treatment planning.



    Fifty-two female Japanese females diagnosed with eating disorders by DSM-III R criteria (24) (avg. age 24.1 years, S.D.=4.45, age range 17-33) attending an outpatient eating disorder group were studied. Participation was voluntary and no remuneration was given. 44 of the patients had anorexia nervosa and bulimia nervosa, 4 had bulimia nervosa, 2 had anorexia, and 2 had eating disorder not otherwise specified. None of the patients had evidence of delusions or thought disorder, active mood disorder, organic mental disorder or mental retardation on clinical exam. The patients averaged 13.6 years of schooling.


    The PBI (8) is a 25-item self rating questionnaire that asks the subject to describe their parents attitudes and behaviours towards them up to the age of 16. Two subscales, "care", and "protection", are scored for the relationship with each parent. The "care" scale measures warmth, empathy, and emotional support. The "protection" scale measures overprotection, control, and intrusiveness. The subject rates each parent on a Likert-type scale, each item scoring 0-3 (Very like-Moderately like- Moderately unlike- Very unlike). "Care" items are reflected by statements like, "was affectionate to me", and "protection" items are reflected by statements like, "tended to baby me".

    While the responses rely on the subjects own recollections, the validity of the instrument has been supported by studies that show that subjects' ratings correlate strongly with ratings of their parents themselves, siblings, and impartial raters (2,8,25). Age, sex, and social class have minimal effect on scores (Y). Also, by administering the PBI to depressed patients and repeating this when they remitted, it has been shown that the care and protection scores were stable over time (25,26).

    The PBI was previously validated in a Japanese population by giving it to 300 high school seniors and their parents (27). Factor analysis yeilded results that were similar to those reported in the original PBI validation studies (8). This study also translated the PBI into Japanese, backtranslating it with minimal differences found.

    The Dissociative Experiences Scale (DES) is a 28-item visual-analog self reporting scale that is a screening instrument for dissociative disorders (28). DES scores over 30 are felt to reflect a high likelihood of post-traumatic stress or multiple personality disorder (MPD)( 29).

    We used data obtained from a modified self-report version of the Dissociative Disorders Interview Schedule (DDIS) (30) to ascertain histories of sexual and physical abuse. The DDIS is a 131-item structured interview that makes DSM III-R diagnoses of all the dissociative disorders, somatization disorder, major depression and borderline personality. It also inquires about a history of substance abuse, child physical and sexual abuse, trance states, schizophrenic symptoms and secondary features of MPD. The DDIS has an overall interrater reliability of 0.68 for the diagnosis of MPD, with a specificity of 100% and a sensitivity of 90%. There is no overall score, results are compared with norms established for MPD in North America. The self-report format was used because Japanese subjects often do not respond directly about sensitive issues in interviews. Detailed analyses of the results of the DDIS given to a Japanese eating disorder cohort are presented in a separate report (31).

    The severity scale of the Bulimic Inventory Test (BITE) (32) was used to measure the severity of bulimic behaviors. Other demographic and patient history data were collected from clinical databases. All the instruments used in this study were in Japanese.


    Subjects signed informed consent and were given the questionnaires to take home for completion. Primary treating clinicials both handed out, as well as collected completed questionnaires. Patients were informed that this was a study looking at various psychologic and childhood experiences, and that the results of the study would in no way affect their treatment. Confidentiality was insured.

    Data Analysis

    Scores were analyzed by correlation testing, t-testing of means, or chi square testing as described in results section. An N for an item reported less than the total N for that group studied reflected a failure of that subject to record an answer for that item or r an "unsure" response.


    (Note: Tables not available on-line).

    PBI scores were compared for those patients who reported sexual abuse or physical abuse to those who did not report a history of that category of abuse (Table 1). Maternal care and both paternal care and protection scores were significantly lower for those subjects who reported physical abuse, but not for those with sexual abuse histories. Maternal protection scores did not differ significantly in either group.

    There were no significant correlations (Fisher's r to z) found between DES scores and PBI scores on any subscale. There were also no significant differences when we looked at the PBI subscale mean scores in those subjects who scored over 30 on the DES (N=7) versus those scoring below 30 on the DES (N=37), and this may have been due in part to a small N. The largest visible difference in means, and the only PBI subscale whose mean score had a trend similar to that seen with any of the abuse comparisons, was that of paternal care (DES<30 mean PBI score of 15.4, S.D.=10.2, versus DES>30 mean PBI score of 10.0, S.D.=10.4), the over 30 DES scorers having lower care similar to that seen in those with physical abuse histories.

    Eating disorder severity as assessed by the BITE was compared to PBI subscale scores as well as to DES scores and no significant correlations (Fisher's r to z) were found.

    While physical abuse history was associated with those subjects who scored over 30 on the DES (4 of 13 with, and 0 of 27 without physical abuse histories, x2=19.4, df=2, p<.0001), presence of sexual abuse history had no significant relation to DES scores. The average DES score of those with histories of physical abuse was 22.8 vs. 7.3 for those without physical abuse histories and this was significantly different (t=4.03, df=35, p<.0003).

    Of the 51 subjects who completed the DDIS, 7 subjects filled the three required criteria for the DSM III-R diagnosis of multiple personality and one subject filled the additional two NIMH criteria for multiple personality. The implications of these findings are presented elsewhere (31).

    The eating disorder patients with histories of either sexual and/or physical abuse did not differ in number of separations experienced as a child compared with those with no abuse history, and both groups depended mostly on their mothers for care. The group with histories of sexual and/or physical abuse however did experience more changes of residence as a child,18 of 24 for the abused group, 11 of 24 for the non-abused group (X2=4.27, df=1, p=.039).


    This is the first report to our knowledge that has compared parental bonding patterns using the PBI in those with, and those without histories of child abuse in a diagnostic sample, and is the first to use the PBI in a psychiatric population in Japan.

    The main positive finding of this study is that patterns of parental bonding significantly differed between those eating disordered patients with histories of child physical abuse compared with those without such histories. Maternal care and Paternal care was significantly lower and paternal protection was significantly greater in those with histories of physical abuse, but not sexual abuse. Maternal protection scores did not differ for any category of abuse.

    The pattern that emerges is that of an affectionless mother and an affectionless/constraining father in those with histories of physical abuse. The bonding patterns measured on the PBI seemed to be useful in deliniating differences between those with and those without physical abuse histories in this cohort. Patterns seen in the family relationships of eating disordered patients may also reflect a reaction to the patient's illness rather than the cause, as well as it is possible that the pattern of parental bonding is a reaction to changes in the patient because of the abuse rather than a primary relationship style.

    One of the reasons why the presence or abscence of sexual abuse histories did not have a significant relationship with paternal bonding as did physical abuse histories, and why physical, but not sexual abuse is related to DES scores over 30, may be that the severity of sexual abuse reported by our subjects did not have the same traumatic impact as that for physical abuse. High dissociation scores have been associated with sexual abuse histories in western studies (29), and it may be that the sexual abuse experienced by the Japanese subjects was milder. This would be consistent with our clinical impressions. Reporting more mild forms of sexual encounters as abuse could potentially dilute any relationship with dissociation or PBI scores.

    While the interpretation of the word "abuse" may have different cultural connotations, our clinical experience is that although sexual abuse seems to occur more frequently in Japan than is acknowledged publicly here, both the frequency and severity of the abuse when it occurs seems to be lower on average than that reported in western studies. We are now examining data on this issue.

    The low incidence of child abuse officially reported in Japan (about 6.6 per 100,000 for children under the age of 12) has been explained by the strong family bond and the high value put on children in Japan (33) though there may be other cultural factors. For example, it is common for Japanese mothers to give up their careers to focus almost exclusively on the upbringing and education of their children. An extreme example of the degree to which the mother-child bond can become fused in Japan is illustrated by the murder-suicide in Japan called "oyako-shinju" (34). This is usually committed by a mother on her young children, and is explained that the mother, under severe family/social stress, finds suicide as the only solution. She cannot imagine her children as being able to live-on without her after her death and psychologically considers them to be a part of herself rather than as separate individuals. Although considered a tragedy, Japanese society in general is sympathetic to the mother in this situation.

    The greater degree of paternal overprotectiveness associated with those with physical abuse histories indicates that, at least for the subgroup with histories of physical abuse, the stereotype of the ineffectual father in eating disorders is inapplicable, and indicates that the father-patient relationship is important for therapy in these patients. It should be noted here as well that abuse history may play an important role in the development of eating disorder symptoms in some patients (14). Because of the differences seen in PBI scores between those with and without abuse histories seen in this study, it seems prudent to take history of abuse into account when interpreting PBI scores in eating disorders, this had not been done in prior studies (1-7).

    The increased frequency of moves experienced by those with histories of sexual and/or physical abuse may reflect more disruption in these families though we did not examine further details in this study.

    Degree of dissociation measured by the DES and PBI scores had no statistically significant correlation. Though a possible trend was seen with high DES scores and low paternal care, this needs to be studied in a larger sample. Because high dissociation is thought to reflect a high likelihood of trauma as a child (29), we hypothesized that parental bonding patterns as measured by the PBI would correlate with DES scores. This was not the case. It could have been that while this was true for some subjects, group heterogeneity may have diluted this out. Alternatively, it may be that the PBI is not sensitive to those bonding patterns that may be associated with high levels of dissociation.

    While it is possible that relationships with parents do not have an effect on and/or are not affected by dissociation, this does not intuitively seem to be the case and requires further study.

    The limitations of this study include a relatively small sample size, cultural and translation considerations in the interpretation of the questions, and problems inherrent in the PBI itself. The PBI relies on the subjects own retrospective recollections rather than actual documentation of paternal bonding patterns. As mentioned above, however, the PBI has been validated in a Japanese sample (27) and studies have shown that subjects' ratings correlate well with those of significant others (2,8,25). While the PBI is limited by the rigid nature of the questions and its limited scope of inquiry, it is more objective than clinical interpretations, though they have the advantage of a fuller and deeper description (1). Additionally, although the self-report format of the DDIS was a modification, it is unclear whether this was a limitation or, within the cultural context used, may have actually strengthened the reliability of the data due to the privacy afforded.

    It seems reasonable to assume that because parental bonding values were less optimal for those with abuse histories the quality of the attachment to the parents is poorer in these cases. This may be one way in which the failure to experience security in one's relationships leads to the development of what Bowlby termed, "anxious attachment" (35). "Anxious attachment", as opposed to the more pejorative term, "overdependency", may predispose to adult psychopathology and may be a useful concept in the interpersonal treatment of these patients (36).

    The results of this study reinforce the importance of bringing child abuse to attention in Japan. The actual incidence of child abuse in Japan does not seem to be reflected by the official statistics, public awareness is minimal, and neither medical nor mental health workers are adequately trained in this area. Consideration of the quality of parental bonding has important implications for therapy, and possibly for social change in Japan. For example, long work-related absences from the home for the father may impact on paternal care. Future studies conducted in a variety of cultures that address the issues presented in this paper can contribute greatly to our understanding of parental relationships and psychopathology.


    The authors thank Drs. Kazunori Nakajima, Ken Murakami, Yoshitomo Takahashi , Yasuhiko Taketomo, Colin Ross, Hiroaki Kumano, and Shinobu Nomura for their helpful comments and support, and Ms. Yoko Kato for technical assistance.


    1.) Palmer RL, Oppenheimer R, Marshall PD: Eating-disordered patients remember their parents: a study using the parental-bonding instrument. International Journal of Eating Disorders 1988; 7:101-106

    2.) Steiger H, Van der Feen J, Goldstein C, Leichner P: Defense styles and parental bonding in eating-disordered women. International Journal of Eating Disorders 1989; 8: 131-140

    3.) Calam R, Waller G, Slade P, Newton T: Eating disorders and percieved relationships with parents. International Journal of Eating Disorders 1990; 9:479-485

    4.) Russel JD, Kopec-Schrader E, Rey JM, Beumont PJV: The parental bonding instrument in adolescent patients with anorexia nervosa. Acta Psychiatr Scand 1992; 86: 236-239

    5.) Gomez J: Learning To Drink: The influence of impaired psychosexual development. J of Psychosom Res 1984; 28:403-410

    6.) Lavik NJ, Clausen SE, Pedersen W: Eating Behavior, drug use, psychopathology and parental bonding in adolescents in Norway. Acta Psychiatr Scand 1991; 84: 387-390

    7.) Pole R, Waller DA, Stewart SM, Parkin-Fiegenbaum: Parental caring versus overprotection in bulimia. International Journal of Eating Disorders 1988; 7:601-606

    8.) Parker G, Tupling H, Brown LB: A parental bonding instrument. British J of Medical Psychology 1979, 52: 1-10

    9.) Oppenheimer R, Howells K, Palmer RL, Chaloner DA: Adverse sexual experience in childhood and clinical eating disorders: a preliminary description. J Psychiatric Res 1985; 19:357-361

    10.) Sloan G, Leichner P: Is there a relationship between sexual abuse or incest and eating disorders? Can J Psychiatry l986; 31:656-660

    11.) Hall RC, Tice L, Beresford TP, Wooley B, Klassen Hall A: Sexual abuse in patients with anorexia nervosa and Bulimia. Psychosomatics 30:73-79

    12.) Shearer SL, Peters CP, Quaytman MS, Ogden RL: Frequency and correlates of childhood sexual and physical abuse histories in adult female borderline inpatients. Am J Psychiatry 1990, 147: 214-216

    13.) Byrne CP, Velamor VR, Cernovsky ZZ, Cortese L, Losztyn S: A comparison of borderline and schizophrenic patients for childhood life events and parent-child relationships. Can J Psychiatry 1990, 35:590-595

    14.) Vanderlinden J, Vandereycken W, Van Dyck R, Vertommen H, Dissociative experiences and trauma in eating disorders. International Journal of Eating Disorders 1993, 13: 187-193

    15.) Torem MS: Covert multiple personality underlying eating disorders. Am J Psychotherapy l990; 44(3): 357-368

    16.) Boon S, Draijer N: Multiple personality disorder in the Netherlands: a clinical investigation of 71 patients. Am J Psychiatry 1993;150:489 494

    17.) Demitrack MA, Putnam FW, Brewerton TD, Brandt HA, Gold PW: Relation of clinical variables to dissociative phenomena in eating disorders. Am J Psychiatry 1990; 147: 1184-1188

    18.) Coons P, Bowman ES: Dissociation and eating. Am J Psychiatry (Letter) 1993; 150:171

    19.) Chandarana MB, Malla MB: Bulimia and dissociative states: a case report. Can J Psychiatry 1989; 34:137-139

    20.) Torem M: Dissociative states presenting as eating disorders. Am J Clin Hypn 1986; 29:137-142

    21.) Moss P: Dissociative disorders clinical update. Carrier Foundation Medical Education Letter 1993; No.176:1-2,7

    22.) Vanderlinden J, VandereyeckenW: The use of hypnosis in the treatment of bulimia nervosa. International J of clinical and experimental hypnosis 1990; 38:101-111

    23.) Minuchin S, Rosman BL, Baker R: Psychosomatic families- anorexia nervosa in context. Cambridge, MA, Harvard University Press, 1978

    24.) American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 3rd ed, Revised, Washington DC American Psychiatric Assn. 1987

    25.) Parker, G: Parental overprotection: a risk factor in psychosocial development? New York, Grune and Stratton, 1983

    26.) Gotlib, IH, Mount JH, Cordy NI, Whiffen VE: Depression and perceptions of early parenting. A longitudinal investigation. Br J Psychiatry 1988; 152:24-27

    27.) Kitamura T, Suzuki T: A validation study of the parental bonding instrument in a Japanese population. The Japanese J of Psychiatry and Neurology 1993; 47:29-36

    28.) Bernstein EM, Putnam FW: Development, reliability, and validity of a dissociation scale. J of Nervous and Mental disease 1986,

    29.) Ross CA, Joshi S, Currie R: Dissociative experiences in the general population. Am J Psychiatry 1990; 147;1547-1552

    30.) Ross CA, Heber S, Norton GR, Anderson D, Anderson G, Barchet P: The dissociative disorders interview schedule: a structured interview. Dissociation 1989; 2(3):169-189

    31.) Berger D, Saito S, Ono Y, Tezuka I, Shirahase J, Kuboki T, Suematsu H: Dissociative symptomatology in an eating disorder cohort in Japan. Acta Psychiatr Scand 1994; 90:274-280

    32.) Henderson M, Freedman CPL: A self-rating scale for bulimia the 'BITE'. Br J of Psych 1987; 150: 18-24

    33.) Takahashi Y: Is multiple personality really rare in Japan? Dissociation l990; 3(2): 57-59

    34.) Takahashi Y, Berger D: Unconscoius processes and suicide in Japan, in Suicide and the Unconscious. Edited by Leenaars A. and Lester D. Northvale, Jason Aronson, 1994 (In Press)

    35.) Bowlby J: Separation: anxiety and anger. New York, Basic Books, 1973

    36.) West M, Rose MS, Sheldon A: Anxious attachment as a determinant of adult psychopathology. J of Nervous and Mental Disease 1993; 181: 422-427

  • Return to Medical Papers


    Acta Psychiatrica Scand, 91:278-282; 1995.

    東京都精神医学総合研究所  バーガー ダグラス  

    東京大学心療内科      久保木 冨房  末松 弘行

    慶応義塾大学        大野 裕  白波瀬 丈一郎  浅井 昌弘

    東京都精神医学総合研究所  斉藤 学  手塚 一郎

    本研究は摂食障害に見られた性的/身体的虐待体験の有無、さらに小児期の虐待体験のあ る患者には多いもである解離症状に対する、親子関係のパターンの格差を検討したもので ある。そこで摂食障害患者にたいする外来集団治療を受けている52例の女性患者に、西 欧で使われている親子関係調査紙「PBI」と解離体験調査紙「DES」と自己記入式解離障 害面接紙「DDIS」修正版(虐待体験も調査し)を実施した。 PBI の二つの下位尺度CAREと PROTECTIONは、父親母親別に評価するものである。

    結果としては、母親のCAREと父親の両下位尺度では、身体的虐待体験のある患者とそうでない患者と比較して、有意差が認められた。性的虐待体験の有無では、母親のCAREだけで有意差が認められた。DES のスコアとPBIのスコアとの間では、各下位尺度には有意な相関が見られなかった。DESのスコアが高い場合(30点以上)と身体的虐待の有無との間には有意差が認められたが、性的虐待との間には有意差が認められなかった。

    以上の所見から、PBIで評価する親子関係のパターンは、小児期虐待体験の有無を区別するために役にだち、さらに親子関係のパターンを研究する場合、この区別を認知しておくべきである。本研究で報告した性的虐待体験は、日本人の患者の場合軽い可能性が考えられるため、PBIのスコアと解離症 状との有意な相関が認められなかった。さらに、解離症状に対する親子関係のパターン区別するにはPBIの有用性が見られなかった。

  • 参考文献に戻る