Evidence based care of eating disorders: beware the glitter of the randomised controlled trial

Objective—The aim of this article is to discuss the practical issues involved in developing evidence based treatment for anorexia nervosa and bulimia nervosa. Method—We discuss the limitations of randomized controlled trials in evaluating treatments for anorexia nervosa. We illustrate how a different strategy is being used in a European Collaborative Research Programme. Results—Many randomized controlled trials of anorexia nervosa have been fraught with difficulties in particular problems with inclusion criteria and compliance. An alternative strategy, that is, a well-designed longitudinal observational study with sophisticated mathematical modelling, has been used in a large-scale German project and has revealed marked variations in models of treatment which are not explained by case mix. Discussion—the randomized control trial may be an inappropriate method to evaluate treatment for anorexia nervosa. Information from longitudinal observational studies will profit both the scientific community and the providers of health care by producing a wealth of new clinical information which can be used to shape service delivery equitably and effectively. A European-wide study, which has the advantages of including a wide variety of health service provision, is currently underway and is an example of such an approach.