Evaluation of Self-esteem Group Therapy for Eating Disorders

Low self-esteem has been established as a strong predictor of onset, maintenance and relapse in eating disorders. For patients with low self-esteem, change can be particularly difficult to achieve, with the potential for prolonged admissions and increased burden of illness to the individual and their family or carers. Significant low self-esteem can contribute to increased striving to control eating, weight and shape to gain a sense of self-worth. Furthermore, it can serve to maintain pervasive negative self- regard and a sense of little or no prospect of recovery. This suggests treatment of low self-esteem is both clinically and financially important. Studies have shown that cognitive behavioural based group therapies can improve self-esteem in adults with eating disorders; however research including adolescents is limited.

This study will contribute to the evidence base for self-esteem interventions in adolescents through evaluating the effectiveness of a new treatment. The self-esteem group intervention evaluated in this study is a novel, manualised group therapy programme designed specifically for adolescent inpatients with eating disorders.

The research will be funded and supported by Newbridge House, inpatient unit for children and adolescents with eating disorders.

Principal Research Question:

Does delivery of a manualised self-esteem group intervention improve self-esteem in adolescent inpatients with eating disorders?

Secondary Research Question:

Are any changes in self-esteem maintained at 4 week follow-up after the group? Does eating disorder psychopathology improve more during the 10 week period for those who complete the self-esteem group compared to those who do not complete the group?

The study aims to generate hypotheses about whether the group intervention is effective in improving self-esteem and/or eating disorder psychopathology.

Research design and methods

Participants:

Participants will be inpatients aged 11-18 who fulfil DSM-V criteria for an eating disorder, meet the framework of inclusion and exclusion criteria and are receiving treatment at Newbridge House child and adolescent inpatient eating disorder unit. A convenience sampling method will be used and participants will be randomly allocated using an Excel formula to either a 'case' group to complete TAU plus self-esteem group or a 'control' group to only complete TAU.

A control group (TAU) was included in order to compare outcomes of those who complete PBI with outcomes of those who do not. This is because the principal aim of the study is to generate evidence of the potential effectiveness of this novel treatment. This is especially necessary given that the research site is an inpatient unit and all participants will be receiving other therapies alongside the treatment being tested.

Forty participants will be recruited in total (twenty in each group). As this is a pilot study no power calculation could be conducted and this sample size was agreed as sufficient to generate data to inform a power calculation for future research. Participants will be identifiable by a unique code created using the letters 'TAU+SE' or 'TAU' followed by a number relating to the order in which they were recruited.

Potential participants will be approached by an Assistant Psychologist when they reach 85% of their healthy weight to introduce the research and seek consent for participation in the study. For patients under the age of 16, informed parental consent and patient assent will be sought. For patients over 16 only informed patient consent will be sought. Patients under 16 will not take part in the study unless both parental consent and patient assent has been received. Patients and parents will be given up to 7 days to decide whether to give consent to participate in the study. Those who consent will then be allocated to the treatment or control group using the excel formula. Once 4 people have been allocated to the treatment group they will begin the self-esteem group.

Design:

To gather evidence for the research questions, a between subjects design (Self-esteem vs no self-esteem) will be used with treatment as the independent variable. The case group will receive TAU as well as the self-esteem group and the control group will just receive TAU. The outcomes from self-report questionnaires completed post-group, 4 week follow-up and at discharge will be compared between the case and control groups.

The dependent variables are self-esteem and eating disorder psychopathology as measured by self-report questionnaires. This experimental design has been chosen as it allows for the comparison of self-esteem vs no self-esteem. The choice of this design was influenced by a desire to gather evidence to generate hypotheses about this novel treatment and to inform a full scale study to test the effectiveness of this treatment for adolescents with eating disorders.

Timetable All stages will take place on site at Newbridge House only.

STAGE 1 (Recruitment) Patients will be approached to take part in the study when they reach 85% of their healthy weight. An assistant psychologist will meet with patients meeting the framework of inclusion and exclusion criteria to read through the research information sheet and request informed written consent. For patients under 16 written, informed consent and patient assent will be sought and parents will be given an information sheet and consent form prior to commencement of the study. For those over the age of 16 written informed consent will be sought only from patients. Patients and parents will be given 7 days to consider their decision to consent to participate in the research. Patients will not be recruited as a participant unless consent has been obtained from the person with parental responsibility and assent has been obtained from the patient, if under 16. If both are not obtained, the patient will not be recruited for the study.

Patients for whom we receive consent will then be randomly allocated to either the case or control group using an Excel formula. Patients who do not consent, or for whom parental consent is not received, will receive TAU which will be identical to that delivered to controls.

STAGE 2 (Week 1) All participants will be asked to complete a set of self-report questionnaires in week 1 (baseline) when they start the group for case participants and when they have consented for the research for control participants. For participants in the case group, the self-esteem group will then commence and take place alongside TAU. For participants in the control group, TAU will continue as normal.

STAGE 3 (Week 6) Stage 3 will take place 6 weeks after completion of baseline measures. All participants will complete a second set of measures.

STAGE 4 (Week 10) Stage 4 will take place 10 weeks after completion of baseline measures. All participants will complete a third set of measures.

STAGE 4 (Date analysis) Stage 5 will take place after all 40 participants have been recruited and completed the study, thus ending the data collection stage. Assistant psychologists on site will then input the raw data from completed outcome measures onto an SPSS spreadsheet. Data will then be analysed and interpreted by a qualified statistician.

STAGE 6 (Report write up) Stage 6 will take place after completion of data analysis. The final report will then be written and submitted for publication by those involved in the research.

Statistical Analysis As a pilot study, the primary aim of the data analysis phase will not be to definitively test whether the self-esteem group is effective or not but to generate evidence of its potential effectiveness if progressed onto a full trial.

In order to do this we will use analysis of covariance (ANCOVA) to estimate the difference in self-reported self-esteem at week 6 and week 10, between conditions, controlling for baseline self-reported self-esteem as a covariate. The difference in self-reported self-esteem scores will be reported (adjusted for baseline levels) within 95% confidence intervals to indicate the potential effectiveness of the self-esteem intervention in comparison to TAU. Primary analysis will be on an intention-to-treat basis. The assumptions or normality and homogeneity of variance will be assessed. Normality assumes that the scores are normally distributed (bell-shaped) and will be assessed using the One-sample Kolmogorov-Smirnov test. Homogeneity of variance assumes that both groups have equal error variances and will be assessed using Levene's Test for the equality of error variances. The relationship between the covariate and the dependent variable should be linear and will be assessed with a scatterplot.

A secondary analysis will explore the difference in self-reported eating disorder psychopathology at discharge between the case and control group.

Drop-out rates between the self-esteem and TAU groups will also be assessed to explore the acceptability of treatment condition. We plan to conduct these analyses using the latest version of IBM SPSS statistics.

Treatment protocol

The intervention will comprise of 6 sessions of group self-esteem therapy which uses a cognitive behavioural approach. The therapy has been manualised into a 76 page treatment protocol and will be administered by Assistant Psychologists employed at Newbridge House receiving weekly supervision from a Clinical Psychologist. The self-esteem group is based on Melanie Fennel's cognitive behavioural model of low self-esteem addressing thoughts, feeling and behaviours. The group aims to improve self-esteem and build positivity.

The sessions are structured and laid out in the following overview:

Week 1, Session 1: What is self-esteem and an introduction to CBT Week 2, Session 2: How might low self-esteem develop? Week 3, Session 3: Negative automatic thoughts Week 4, Session 4: Challenging negative automatic thoughts Week 5, Session 5: Creating new coping strategies Week 6, Session 6: Building positivity