Psychotherapy Outcome Research with Bulimia nervosa

Abstract

Recent psychotherapy studies for bulimia nervosa are reviewed and it is concluded that many interventions lead to statistically and clinically significant reductions in the frequency of bingeing and vomiting and other associated symptoms; however, the results are more variable if success is defined as abstinence from symptomatic behavior. Treatment and patient variables which seem to be associated with positive response to treatment are reviewed and serious methodological concerns are raised. Carefully controlled dismantling trials are recommended in order to isolate factors which are critical for symptom change.

Does early intervention improve the prognosis in anorexia nervosa? A systematic review of the treatment-outcome literature

“…On the basis of the lack of prognostic value of the duration of the illness in these homogenous studies no far-reaching inferences should be made (Garner, 1987). The only conclusion we can draw from the studies conducted in adolescent clinics is that a few months delay in admission does not appear to influence the outcome of treatment-which is in line with clinical experience.…”

“…(1) Small sample sizes often produce inadequate power (Garner, 1987;Kazdin, 1986). In accordance with Hsu (1980), a minimum of 15 patients was therefore required.…”

“…The nonsignificant relation found between the duration of the illness and treatment outcome could be due to the limited variability in the predictor variable duration of illness. While this ''low within-group variability'' argument, which was mentioned earlier by Garner (1987) and Szmukler and Russell (1986) as an explanation for the frequent lack of relation between prognostic factors and treatment outcome in eating disorder research, seems plausible (Kazdin, 1986), it has yet to be explicitly investigated.…”

“…As at the time no more than one study used some form of randomization (Gowers, Norton, Halek, & Crisp, 1994), this problem holds for nearly all of the studies of the treatment outcome of AN. The only way to exclude such alternative explanations in nonrandomized studies is to measure the influence of potentially confounding variables and then control for these in multivariate analyses (Garner, 1987;Herzog, Rathner et al, 1992;Steinhausen & Glanville, 1983a;Szmukler & Russell, 1986;Yates, 1990). It should be noted that such multivariate statistical control is impossible when patients with previous treatment are not considered separately as the unique influences of denial, comorbidity, negative selection, treatment failures, and chronicity are inextricably interwoven in such cases (Goldberg et al, 1980).…”

Therapy Outcome of Two Treatment Models for Bulimia nervosa: Preliminary Results of a Controlled Study

“…Beyond the reduction of core symptoms after either therapy, different modification patterns are expected due to differing emphases in the two therapy groups. So far there is little empirical proof in the literature on bulimia nervosa for either the efficacy of the conflict-centered inpatient analytic therapy or the outpatient systemic one [ 1,2], This report is based on the initial results of a study in which patients awaiting treatment served as controls.…”