How effective is family therapy for the treatment of anorexia nervosa?

Guidelines for the treatment of patients with eating disorders have recently been published in many countries (e.g.Wilson&Shafran, 2005). These guidelines should be based on scientific evidence, hopefully, evidence from randomised controlled trials, the gold standard of medical research. Such clinical trials are ‘indispensable ordeals’ for testing the validity of therapeutic hypothesis (Frederickson, 1980). Once guidelines have been published, cliniciansmay take for granted that the interventions that have been recommended have a firm scientific basis and that they thereforedonot need to consult the original data that supports the recommendations. Recently published guidelines states that: ‘Of the three recommendations that are a priority for implementation in anorexia nervosa, the strongest is that children and adolescents should be offered family interventions’ (Wilson & Shafran, 2005). However, to cite a recent review: ‘Surprisingly little systematic research on the efficacy of family therapy for eating disorders exists despite its common clinical use’ (Lock & le Grange, 2005). Here we review that research, considering only evidence obtained in randomised controlled trials. In an often cited study on the effect of family therapy and individual therapy on anorexic patients, Russell, Szmukler, Dare, and Eisler (1987) reported that the overall outcomewas poor in 54 participants. However, they did report an improvement inmildly affected individuals (mean age1⁄4 16.6 years and mean duration of illness1⁄4 1.2 years), but they noted that treatment was not effective for either adults or adolescents who had the disorder for a prolonged period of time. Thus, in this study there was a subgroup of only 10 young mildly ill patients who improved with family therapy compared to 11 patients who received individualised therapy. In a 5-year follow-up of these results, Eisler, Dare, Russell, Szmukler, le Grange, and Dodge (1997) found that the outcome had improved in all participants, regardless of treatment, and they concluded that: ‘Much of the improvements found at a 5-year follow-up can be attributed to the natural outcome of the illness’. In other word, neither family therapy nor individual therapy had much of an effect on the long-term outcome of this disorder. In second study by this group, Eisler, Dare,Hodes, Russell, Dodge, and le Grange (2000) treated even younger patients who were even less ill. These patients were on average 15.5 years old and they had been ill for less than a year. Only 15 of 40 of these mildly affected patients achieved what they described as their good outcome. Specifically, at the termination of treatment, patients with a ‘good’ outcome achieved only 87% of normal weight, only 44% of the postpubertal girls were menstruating, they had no change on the Psychosexual Scale and they showed only modest gains in self-esteem and obsessional thoughts. To illustrate their difficulty in treating seriously ill patients, 8 of 11 patients who had any previous treatment had a poor outcome. Moreover, the longer their patients had been ill, or the more emaciated they were, the less likely they were to improve. It is also quite striking that the improvements in the anorexic symptoms among all their patients were accompanied by a significant increase in the bulimic symptoms, raising the