This study will compare the medical and cost effectiveness of a stepped approach, including self-help and drug therapies, to the current best available care for bulimia nervosa.

The eating disorder bulimia nervosa (BN) is estimated to exist in 1 % to 2 % of women from the ages of 15 to 30, and bulimic behaviors occur in many more who do not develop the disorder. Characteristic symptoms include binge eating and self-induced vomiting. Research indicates that cognitive behavioral therapy (CBT), paired with use of the antidepressant fluoxetine, is the treatment of choice for BN. However, a survey of doctoral level psychologists dealing with eating disorders found that 78% of clinicians had no training in CBT for eating disorders, with 72% also untrained in the alternative, interpersonal therapy. Self-help manuals based on CBT have been developed as a treatment option for those without access to a clinician trained in CBT for eating disorders. Studies of these manuals suggest that self-guided therapy is as effective as CBT for some individuals. Additionally, although little research has examined cost effectiveness for treatment of eating disorders, a self-help approach may be more cost effective than CBT. This study will test whether self-help therapy, in the context of a stepped treatment plan involving check-ins with a clinician and possible drug treatment, can be an effective alternative, medically and economically, to traditional CBT.

Participants at four treatment sites will be screened for diagnosis, health, and presence of other conditions through assessments in standardized questionnaires, clinical interviews, and a blood test. Participants who pass the screening will be randomly assigned to one of two treatment sequences. In the first sequence, participants will be given twenty 50-minute sessions of individual CBT treatment over 18 weeks, focusing on altering eating behaviors. If participants do not have an adequate initial response to treatment by session six, they will be offered a trial of fluoxetine in addition to their continued CBT treatment. If they do not meet response criteria at the end of the 18 weeks of CBT treatment, they will again be given the option of taking or continuing to take fluoxetine until the termination of the study. The second sequence will begin with a specially designed self-help program aimed at teaching the same skills that will be taught in CBT sessions. Participants will have nine 30-minute checkup sessions with a therapist over 18 weeks. Participants who do not respond adequately to the treatment after 10 weeks will be given the option of taking fluoxetine. Participants without sufficient recovery at the end of 18 weeks will be given the option of undergoing the first treatment sequence.

Assessments of costs and health will be made at five points: the beginning of the study, and at the study visits on Week 10, Week 18, Week 36, and Week 62. Multiple assessments will be used to test health, including eating disorder symptoms, presence of other psychopathology and personality factors, social and interpersonal functioning, and quality of life. Eating disorder symptoms will be evaluated through clinical interview, questionnaires, and weight fluctuations. Treatment variables will also be assessed, with particular interest in the measure of patient knowledge of CBT techniques, to determine how much information from the self-help manual is used and absorbed. For the cost benefit analysis, assessments will include cost of health care visits, medications, treatments, and time lost due to illness. Data on cost will be collected at study visits, except for time lost, which will be measured through a questionnaire completed by a family member once at baseline and once after 18 weeks of treatment.

Inclusion Criteria:

• Diagnosis of bulimia nervosa

Exclusion Criteria:

• Weight less than 85% of ideal
• Current substance abuse or dependence
• Pregnant
• Currently receiving psychotherapy