A naturalistic study of psychotherapy for bulimia nervosa, part 1: comorbidity and therapeutic outcome.

Data from naturalistic samples provide an important complement to findings from randomized trials of psychotherapy. A random national sample of US clinicians provided data on 145 completed treatments of patients with bulimic symptoms. Treatment in the community was substantially longer than treatment prescribed in manuals, with a mean length of cognitive-behavioral therapy of 69 sessions and significantly longer for eclectic and psychodynamic therapies. Most patients treated in the community had substantial comorbidity, and this comorbidity was associated with longer treatments and poorer outcome. Using four common exclusion criteria from randomized controlled trials for bulimia nervosa, approximately 40% of the naturalistic sample would have been excluded from randomized controlled trials. These patients showed higher pretreatment severity and required longer treatments to achieve positive outcomes relative to patients who did not meet these exclusion criteria.