Implications for the treatment of bulimia nervosa: A meta-analysis of efficacy trials and a naturalistic study of treatment in the community.

Abstract

Prior meta-analyses of the clinical trials of treatments for Bulimia Nervosa have concluded that short-term cognitive-behavioral therapy (CBT) is the treatment of choice for all patients. This dissertation first presents a multidimensional meta-analysis of the same clinical trial data previously analyzed, using additional variables bearing on generalizability and outcome. The data suggest that short-term treatments do result in substantial improvement: However, almost forty percent of the patients who applied for treatment were excluded from the clinical trials; over half of the patients who entered treatment did not recover; and the average patient who completed treatment showed high symptom levels at the post-treatment timepoint. The selection procedure for inclusion in these studies raises questions about representativeness and generalizability of these clinical samples, while the outcome results raise questions about the efficacy of short-term CBT as tested in clinical trials. To assess treatment techniques and outcomes in a less selected sample, this dissertation next presents original data from a naturalistic study of treatment of patients with bulimic symptoms in the community. The clinician-report data suggest that treatments in the community (unconstrained by the limitations of clinical trials) are of much longer duration than the treatments provided in the manuals for clinical trials, and address a much more varied population. Comorbid axis I and axis II diagnoses were extremely common in the naturalistic sample, and it appeared that at least forty percent of the patients with bulimic symptoms treated in the community would have been excluded from clinical trials under four common exclusion criteria. The data suggested that there were three subtypes of bulimia represented in the sample: a High Functioning/Perfectionistic subtype; a Dysregulated/Undercontrolled subtype; and a Constricted/Overcontrolled subtype. These subtypes showed different patterns of comorbidity and different treatment outcomes. The Dysregulated patients had the most comorbidity, the longest treatments, and the least successful outcomes. In addition, the therapists reported using different therapeutic techniques with the different subgroups: both self-declared CBT and Psychodynamic therapists used more Psychodynamic interventions with the Dysregulated patients. Both CBT and Psychodynamic approaches appeared to have specific effects on treatment outcome: the use of CBT was correlated with shorter overall therapies and shorter times to improvement in eating symptoms, while the use of Psychodynamic psychotherapy was correlated with more overall improvement (particularly among the Constricted and Dysregulated subgroups). The two studies taken together imply that previous claims on behalf of short-term CBT as the treatment of choice are in need of clarification.