Unresolved attachment and alliance in three treatments for borderline personality disorder

Open Access
Clouthier, Tracy Lynn
Graduate Program:
Master of Science
Document Type:
Master Thesis
Date of Defense:
April 10, 2015
Committee Members:
  • Kenneth Levy, Thesis Advisor
  • psychotherapy
  • attachment
  • alliance
  • borderline personality disorder
Therapeutic alliance is an important predictor of outcome in psychotherapy (Castonguay et al., 2006; Martin et al., 2000). However, patient differences in capacity to form an alliance likely influence this relationship between alliance and outcome (Clarkin & Levy, 2003). Levy and colleagues (2011) found that attachment security predicts better outcome. In a follow-up meta-analysis, they found that self-reported insecure attachment predicted poorer alliance (Bernecker et al., 2014). However, few studies have examined the role of unresolved attachment organization in predicting alliance and outcome in psychotherapy. Unresolved attachment organization, which indicates the presence of unresolved loss or trauma, has been associated with higher rates of psychopathology and other negative outcomes (Riggs et al., 2007), and may predict difficulty effectively engaging in treatment (Korfmacher et al., 1997). In the present study we examine how unresolved attachment impacts early alliance in three treatments for BPD. Ninety patients from a randomized controlled trial (RCT) comparing transference-focused psychotherapy (TFP), dialectical behavior therapy (DBT), and supportive psychotherapy (SPT) (Clarkin et al., 2007) were assessed for attachment organization using the Adult Attachment Interview (AAI). Patients were classified as unresolved with regard to loss or trauma if they exhibited disorganization, dissociation or lapses in monitoring their speech or reasoning when discussing past trauma and/or losses during the AAI. For each patient, three early sessions were coded for alliance by trained raters with the Working Alliance Inventory (WAI): two at about one month into treatment, and a third about three months into treatment. Ratings from the first two sessions were averaged into a single time point in order to control for any discrepancy resulting from a focus on setting the therapeutic frame. A two-way (unresolved organization x treatment group) ANOVA was used in order to assess whether there is any interaction between unresolved attachment organization and treatment orientation in predicting the quality of the therapeutic alliance. The main effects of treatment and unresolved attachment on alliance were nonsignificant at both time points. At the first time point, there was a significant interaction between unresolved attachment and treatment group. There was also a trend-level interaction between these variables at the second time point. Findings from this study could potentially have both prognostic implications regarding the risk presented by unresolved attachment and prescriptive value for patient treatment matching to enhance alliance for those with unresolved trauma or loss.