Comparing CBT-T to CBT-E in Non-low-weight Adults With Eating Disorders

Overview

Eating disorders are a difficult to treat illness with significant psychological and physical sequelae. Cognitive behavioural therapy (CBT) has been the most researched and supported intervention for eating disorders. A particular version of CBT for eating disorders, CBT-E (Fairburn, 2008), has been the focus of much research over the past decade. Despite promising results from initial CBT-E trials these findings have not always replicated well and evidence points to high drop-out in real-world settings. Further, CBT-E is a resource-intensive intervention, which may contribute to poorer access to care. In an attempt to overcome some of these barriers related to CBT-E, Waller and colleagues (2018) recently developed a brief (10 session) version of CBT for non-low-weight eating disorders (e.g., bulimia nervosa and binge eating disorder), referred to as CBT-T. Preliminary evidence from a case series of adult patients suggests that CBT-T has similar efficacy to CBT-E with low rates of drop-out. However, further evaluation of this brief treatment is needed, including direct comparisons with CBT-E. Indeed, given that no comparison group was included in the initial case series, it is unknown whether either CBT-T or CBT-E may be superior to the other. Thus, the aim of the current project is to examine CBT-T's efficacy in comparison to CBT-E as it has been implemented at the eating disorders service at London Health Science's Centre, and to determine whether either intervention is superior based on treatment outcome and treatment drop-out. Non-low-weight individuals with eating disorders assessed at the Adult Eating Disorders Service will be eligible to participate in the study. The principle investigator is Dr. Philip Masson, Ph.D., C. Psych., 519-685-8500 ext. 74866.

Design: As CBT-T is anticipated to achieve faster improvement, due to its brief nature, than CBT-E this research study is not utilizing a non-inferiority design. That is, differences between the two interventions are anticipated, at least initially. Over time, the two interventions may become similar in some aspects (final remission rate for completers) but not others (final remission rate utilizing intention to treat). Hypotheses: CBT-T will produce higher rates of symptom abstinence initially as the treatment takes approximately half as long as CBT-E and then at post-treatment follow-ups that effect will disappear using completer analysis. However, as drop-out is expected to be lower in CBT-T it is expected that CBT-T will consistently produce superior results to CBT-E with intention to treat analysis. Recruitment: Individuals who have been assessed at the Adult Eating Disorders Service and meet eligibility criteria will be informed by the assessor that they are eligible to participate in a trial examining a new treatment. The assessor will provide the letter of information for the study and the study will be explained to the participants and any questions the participants have about the study will be addressed. The assessor will inform the staff psychometrist at the service that the participant is eligible to participate in the study. Participants will be asked to participate in the study once the participants name has reached the top of the treatment waitlist and a clinician is able to see them. Once the patient's name reaches the top of the treatment wait-list the participant will meet with the psychometrist in preparation for treatment, as part of routine procedure, and asked to participate in the study. Due to the significant waiting list (currently 5-6 months) as well as the tendency for some clients to end up not accessing care, study participation will be sought when it is most relevant. At the beginning of client's appointments with the staff psychometrist, patients will be reminded of the study, provided information about the study and asked if to participate in the study. If they indicate interest in participating, then the psychometrist will thoroughly review the letter of information with patients. If patients agree to participate and sign the consent form, study measures will be given. The psychometrist will then inform the assigned therapist that the client wishes to participate in the study. To ensure the psychometrist is not aware of randomization, Philip Masson (study PI) will randomize the client to one of the two interventions using the website random.org. Philip Masson will inform the therapist of the randomization, and the therapist will then proceed with the appropriate treatment intervention. The PI will then keep track of randomization assignment in a separate excel file that the psychometrist does not have access to. If clients do not wish to participate in the study, clients will be assessed using the clinic's standard program evaluation protocol and will receive treatment as usual (CBT-E). Once individuals are participating in the study only the assigned treatment will be received and this process will be clearly explained to participants. Participants can withdraw from the study at any time at which point their data collection would stop; however, participants would continue to receive their assigned treatment. If participants wish to end treatment, then participants will be able to reapply to the service after a 2-month waiting period, which is standard practice for the service. Interventions: CBT-E is a transdiagnostic eating disorders treatment based on the principles of cognitive behavioural therapy which has been developed over the past 40 years. The treatment focusses on tracking and modifying eating behaviours, providing psychoeducation, addressing overvaluation of weight and shape, body image, food restriction and restraint, emotion regulation, and relapse prevention. For non-low-weight individuals, the treatment is designed as a 21 session treatment (one orientation meeting and 20 treatment sessions). The adult eating disorders service has been utilizing this treatment approach on an outpatient basis for the past two years. All staff members have reviewed the reading material associated with the treatment, participated in online training for CBT-E and participate in weekly peer supervisions sessions where staff members support one another in utilization of the treatment. Therapists are encouraged to review relevant sections of the therapist manual as needed. The treatment is adhered to the manual as close as possible except for one major exception. In CBT-E the treatment is designed to be delivered twice weekly for the first 4 hours. Instead, the service provides weekly sessions as the patients and staff found it too difficult to make twice weekly sessions fit into their schedule. After approximately 17 weekly sessions then sessions drop down to every other week. CBT-T is a 10-session treatment based on cognitive behavioural therapy. It was developed based on clinical experience as well as core components of evidence-based versions of CBT for eating disorders, including CBT-E. The aim of the intervention is to provide only the hypothesized critical elements of CBT for eating disorders so that the interventions can be provided quicker. As rapid response to treatment in CBT for eating disorders has been shown to be a key predictor of outcome, it is possible that a briefer intervention may not be necessarily less effective. Further, briefer interventions may result in reduced drop-out which would allow more clients to fully access treatment. The key elements of this intervention are similar to CBT-E except there is a less explicit focus on emotion regulation and the psychoeducation is provided in the moment as opposed to being provided as a separate element of treatment. Further, in CBT-T treatment is initially only offered for four sessions and then extended only if the patient is actively engaging in treatment, which is similar to the session 8 progress review in CBT-E but occurs earlier and is framed differently. The therapy is delivered using a session by session checklist of core tasks to facilitate adherence. The clinicians at the service will receive training in CBT-T in the form of an 8-hour workshop delivered by one of the creators of the intervention, Glenn Waller. Further, the treatment manual will also be provided to all clinicians and read prior to the workshop. The clinicians will also participate in group supervision for one hour per week focusing on effective delivery of CBT-T. Data Analysis: Data will be analyzed using SPSS (version 24). Treatment groups will be compared on baseline characteristics (e.g., eating disorder symptoms) using independent samples t-tests and chi-square for categorical measures. Percentage scores will be calculated to determine attrition rates at each stage of treatment and chi-square analyses will be used to compare attrition/drop-out between the two treatment conditions. A two-level hierarchical linear model will be used to examine the influence of the intervention on change in symptoms over time. Level 1 (within-subjects) will include repeated measures of the dependent variable (e.g., EDE-Q scores, ED symptoms) and level 2 (between-subjects) will include treatment condition (CBT-T or CBT-E) and the interaction between condition and time. Separate models will be conducted for different outcome variables of interest. Both an intention to treat and completer analyses will be performed. For the intention to treat analysis multiple imputation will be utilized to handle the resultant missing data. Power analyses: In accordance with recommendations for pilot RCTs by Rounsaville, Carroll, \& Onken (2001), the investigators aim to have at least 20 participants per group (after attrition). Based on power analyses for repeated measures analysis of variance (ANOVA) for between-within group differences using G\*power 3, a total sample size of 40 would yield \>80% power to detect medium effects (f2 \>.20), which represents a conservative approach, given increased power of multilevel modeling over traditional repeated measures ANOVA. Based on prior studies of CBT-T, the investigators anticipate an overall attrition rate of approximately 30%; thus, they propose to enroll 60 participants (30 per treatment condition) to detect an attrition-adjusted medium effect.