اتحاد درمانی در بهبود رفتار درمانی شناختی برای بولیمیا: احتمالا ضروری است اما قطعا کافی نیست

The present paper assessed therapeutic alliance over the course of Enhanced Cognitive Behavioural Therapy (CBT-E) in a community-based sample of 112 patients with a diagnosis of bulimia nervosa (BN) or atypical BN. Temporal assessment of alliance was conducted at three time points (the start, middle and end of treatment) and the relationship between alliance and treatment retention and outcome was explored. Results indicated that the alliance between patient and therapist was strong at all stages of CBT-E, and even improved in the early stages of treatment when behaviour change was initiated (weekly in-session weighing, establishing regular eating, and ceasing binge-eating and compensatory behaviours). The present study found no evidence that alliance was related to treatment retention or outcomes, or that symptom severity or problematic interpersonal styles interacted with alliance to influence outcomes. Alliance was also unrelated to baseline emotional or interpersonal difficulties. The study provides no evidence that alliance has clinical utility for the prediction of treatment retention or outcome in CBT-E for BN, even for individuals with severe symptoms or problematic interpersonal styles. Early symptom change was the best predictor of outcome in CBT-E. Further research is needed to determine whether these results are generalizable to patients with anorexia nervosa.

Manual-based treatments, such as Enhanced Cognitive Behavioural Therapy (CBT-E), are amongst the most effective treatments for bulimia nervosa (BN) currently available. Yet a prevailing view is that treatment outcome is related to individual therapist differences over and above therapeutic approach (e.g., Luborsky et al., 1986 and Messer and Wampold, 2002) and that manual-based treatment approaches are less caring, less intuitive, less authentic, and even inappropriate for ‘real-world clients’ (Addis & Krasnow, 2000). Empirical research provides no evidence that therapeutic alliance is adversely affected by implementing manual-based treatments for BN, with patients rating the alliance favourably in randomised controlled trials (RCTs; Loeb et al., 2005) and naturalistic clinic settings (Waller, Evans, & Stringer, 2012). Therapeutic alliance (hitherto referred to as “alliance”) can be defined as establishing shared goals between patient and therapist; accepting the tasks that each needs to perform; and the attachment bond between the patient and therapist (Bordin, 1979). Beyond eating disorders, the finding of a relationship between alliance and psychotherapy outcomes prevails across different measures and definitions of alliance and across different interventions (Horvath and Symonds, 1991 and Krupnick et al., 1996). The association between alliance and treatment outcomes has been observed to be small but robust, with meta-analyses reporting effect sizes ranging from 0.22 to 0.26 (weighted correlation coefficient e.g., Horvath & Bedi, 2002). There is mixed evidence regarding the relationship between alliance and treatment outcome in eating disorders. Alliance has been shown to be associated with changes in eating disorder symptoms in anorexia nervosa (AN; e.g., Isserlin and Couturier, 2012 and Pereira et al., 2006) and in BN (e.g., Constantino et al., 2005 and Treasure et al., 1999). In a large RCT, stronger early (session 4) and mid-treatment (session 12) alliance was associated with fewer purge episodes at the end of CBT for BN, after accounting for baseline purge frequency (Constantino et al., 2005). Other studies fail to find that alliance is associated with change in eating disorder symptoms (e.g., Brown et al., 2013, Loeb et al., 2005 and Waller et al., 2012). In a clinical trial comparing interpersonal psychotherapy (IPT) and CBT for BN, alliance at sessions 6, 10, or 18 failed to predict post-treatment purge frequency in either treatment condition, after accounting for baseline purge frequency (Loeb et al., 2005). There is also evidence that improvements in eating disorder symptoms may precede improvements in alliance ratings. In CBT for AN, Brown et al. (2013) observed that early weight gain preceded improvements in alliance ratings. In IPT for BN, Loeb et al. (2005) observed that reductions in episodes of vomiting preceded improvements in alliance ratings. In BN, Wilson et al. (1999) found that higher alliance was associated with greater likelihood of achieving full remission across four treatment conditions (CBT, supportive psychotherapy, anti-depressant medication, and placebo), however temporal analysis indicated that early symptom improvement was more reliably associated with subsequent higher alliance ratings than vice-versa. These findings invite the conclusion that symptom improvement might drive more positive ratings of alliance and highlight the importance of considering temporal factors (particularly early symptom change),1 when considering the relationship between alliance and treatment outcome in eating disorders. Few studies have investigated the relationship between alliance and treatment retention in manual-based treatments for eating disorders. Brown et al. (2013) found no evidence that alliance at session 6 predicted retention in CBT for AN. Carter et al. (2012) failed to find an association between early alliance and retention in CBT-E with a transdiagnostic sample. There is no research on the relationship between alliance and retention in CBT-E for BN. A limitation of existing studies is that the earliest measure of alliance occurs well after symptom change has commenced. In several studies the earliest measure of alliance has been at session 6 (e.g., Brown et al., 2013, Loeb et al., 2005 and Waller et al., 2012), and Loeb et al. (2005) noted that 74% of the change in BN symptoms in CBT occurred prior to session 6. This leaves the possibility that an earlier measure of alliance might better predict treatment outcome and retention, with Waller et al. (2012) recommending that future studies measure the alliance from the earliest time point in therapy. One study found no evidence that alliance at session two predicted drop-out from CBT-E in a transdiagnostic sample of eating disorder patients (Carter et al., 2012). This study did not explore the relationship between alliance and treatment outcome or temporal patterns of alliance over the course of CBT-E, and conclusions are limited due to the heterogeneous nature of the clinical sample. The importance of developing a strong alliance in terms of treatment outcomes may be influenced (i.e., moderated) by individual patient characteristics. For instance, a strong alliance might be more difficult to establish for patients with a history of problematic interpersonal relationships, and yet paradoxically having a strong alliance might be particularly important for optimising outcomes for these individuals. A strong alliance may increase engagement for such patients and thus provide a powerful therapeutic context for identifying and modifying problematic behaviours. Additionally, a strong alliance might be particularly important for patients with more severe symptoms as they endure the emotional rigours of engaging in behavioural change. In a transdiagnostic sample of eating disorder patients, Waller et al. (2012) found that patients with higher levels of emotional distress (anxiety, depression, and interpersonal sensitivity) were more likely to report that the goals of therapy were less well shared with the therapist at session 6. Patients with higher scores on psychoticism, depression, and interpersonal sensitivity scales also reported less positive attachments with their therapist at session 6. Constantino et al. (2005) observed that patients with more interpersonal difficulties at baseline had poorer alliance in the middle of treatment in IPT but not CBT. When exploring the relationship between alliance and outcomes in manual-based treatments for BN, it is therefore important to consider the potential influences of factors such as anxiety, depression, and interpersonal difficulties. Importantly, the question of whether problematic interpersonal styles and symptom severity interact with therapeutic alliance to predict symptom improvement is yet to be answered. The current paper examined alliance in a community-based sample of patients with BN or atypical BN participating in individual CBT-E. Alliance was measured at the start (session two), middle (week 10), and end of treatment. The paper evaluated patient ratings of alliance over the course of CBT-E and examined the relationship between alliance and treatment outcome and retention. No studies have investigated very early alliance in BN or alliance in CBT-E for BN. The first hypothesis was that alliance would be high throughout treatment, as established in previous studies of manual-based treatments for eating disorders. The second hypothesis was that alliance would be related to treatment retention and baseline anxiety, depression, and interpersonal difficulties. Finally, we examined whether problematic interpersonal styles or symptom severity, and their interaction with alliance, were associated with outcome. The third hypothesis was that alliance would be particularly important for individuals with higher levels of interpersonal problems and with more severe eating disorder symptoms.