Psychological treatment for anorexia nervosa: a meta-analysis of standardized mean change

For the German treatment guidelines for eating disorders, the literature on psychological treatment of anorexia nervosa (AN) was reviewed systematically. As a common meta-analysis of randomized clinical trials proved to be impossible, a review of all available clinical trials was conducted, statistically integrating standardized mean change scores. Research questions comprised differential effects of therapeutic techniques and settings as well as determining which weight gains could be expected.

After an extensive literature search, studies were selected, rated by 3 independent raters. Weight gain as the main outcome criterion was transformed into standardized mean change scores. Effect sizes were checked for homogeneity.

57 studies containing 84 treatment arms and 2,273 patients could be integrated. Studies differed considerably in quality. The strongest bias identified was reporting selectively on completers or failures, versus intention-to-treat samples. No significant differences between effect sizes could be identified concerning treatment setting, technique or patient characteristics. If treatment time is taken into account, inpatient treatment produced a faster weight gain than outpatient treatment.

The study describes weight gains which can be reached in outpatient and inpatient settings. It yielded no salient results speaking for a certain therapy technique, setting or procedure. Treatment guidelines for psychological treatment of AN still have to rely on lower level evidence.