Variability in remission in family therapy for anorexia nervosa.

Abstract

The evolution toward more stringent conceptualizations of remission in family therapy for adolescent anorexia nervosa (AN) has, with time, introduced variability in outcomes across randomized controlled trials (RCTs). An examination of remission across the history of research on family therapy for AN shows that earlier studies adopted lenient definitions and generally yielded higher rates of remission than studies of the past decade that have used stricter definitions of remission. In this study, we investigate the reactivity of remission rates to the application of different definitions of remission used within the family therapy for AN literature, within a single RCT data set. We conducted a secondary analysis of data from a single-site RCT which compared the relative efficacy of two formats of family therapy in a sample of 106 Australian adolescents with AN. Using end-of-treatment data, we compared remission rates using 11 definitions of remission that have been used in studies of family therapy for AN spanning more than three decades. We found wide variability in remission rates (21.7-87.7%; Cochran's Q χ2 (10, N = 106) = 303.55, p = .000], depending on which definition of remission was applied. As expected, more lenient criteria produced higher remission rates than more stringent definitions. Applying different criteria of remission to a single data set illustrates the impact of changing how remission is defined. Failure to consider the greater stringency of remission criteria in recent studies could result in false inferences concerning the efficacy of family therapy for AN over time.