Insecure Internal Working Models (IWMs) of women with Eating Disorders with and without Borderline Personality Disorders

Introduction. Many studies supports the hypothesis that women with eating disorders (ED) show dismissing Internal Working Models (IWMs). However, others revealed that they often display both dismissing and preoccupied IWMs, or a prevalence of the latter. Moreover, a relevant presence of unresolved with respect to loss and/or trauma emerged among ED patients. In addition, no studies up to now have considered the issue of co-morbidity of ED with Borderline Personality Disorder (BPD). Objectives. Ours goals have been twofold. First we examined the IWMs of women with ED compared with a matched non-clinical group. Second, we explored the differences of IWMs among ED patients with and without BPD’s diagnosis. Methods. Clinical participants were 50 patients with ED, all between 16 and 25 years, and without any concurrent severe psychiatric disturbance on Axis-I (e.g. psychosis, etc). Moreover, 50 non clinical participants were recruited from community and matched for sex, age, years of education. The Adult Attachment Interview (AAI) was used to assess the attachment’s IWMs. All AAI transcripts were analyzed by two expert and blinded coders on the AAI’s coding system, yielding five global categories: free-autonomous, insecure-dismissing, insecure-preoccupied, unresolved and cannot classify. Results. An over-representation of insecure-dismissing attachment IWMs was revealed among eating disordered women. Moreover patients who had a double diagnosis of ED and BPD showed higher scores on involving anger and unresolved loss compared with those with only ED. Discussion. These results seem to be in line with Dozier’s hypothesis suggesting that individuals with psychological disorders leading to an outward focus (“externalizing” perspective) –such as ED– would go together with a more dismissive stance to their own attachment experiences and a minimization of their attachment needs. Further longitudinal research, using a larger community, both at risk and clinical samples, could be useful to explore the causal process direction