Twenty-six studies were included in the review: two randomised controlled trials (RCTs; 195 patients); 16 non-randomised controlled studies (1,306 patients); and eight non-controlled studies (279 patients). Sample sizes ranged from eight to 120 patients. Completion rates in the studies varied from 30.8% to 100%.
All eight uncontrolled trials of self-help for bulimia nervosa or EDNOS reported reductions in binge eating (range 33% to 85%) and self-induced vomiting (17% to 80%). Studies reported abstinence rates of between 12.8% and 47% for binge eating, 27% and 65.2% for vomiting and 26.8% and 50% for binge-eating and vomiting.
Seven out of the eight studies that compared self-help with waiting list control for bulimia nervosa or binge-eating disorder reported a significant advantage for self-help. Improvement rates ranged from 25% to 87% for self-help and 6% to 19% for waiting list control; abstinence rates ranged from 6% to 53% for self-help and 0% to 13% for waiting list control.
Two of eight studies that compared self-help to another active comparator found significant differences in favour of guided self-help in comparison with pure self-help (one study) and standard management (one study). Six studies found no significant differences between study groups.
Three studies that compared a combination of self-help with medication for bulimia nervosa or binge-eating disorder reported mixed findings.
Two studies that compared a combination of self-help and CBT with CBT alone reported no significant differences between the two intervention types.