Screening for eating disorders and high-risk behavior: caution.

Abstract

OBJECTIVE: The current study reviews the state of eating disorder screens. METHODS: Screens were classified by their purported screening function: identification of cases with (a) anorexia nervosa only; (b) bulimia nervosa only; (c) eating disorders in general; (d) partial syndrome, eating disorder not otherwise specified (EDNOS), or subclinical; (e) not a-d but at high risk. Information is presented on development, psychometric properties, and external validation (e.g., sensitivity, specificity, positive predictive values, and negative predictive values). RESULTS: Screens differ widely with regard to objective, psychometric properties and the validation methodology used. Most screens that identify cases are not appropriate for the identification of at-risk behaviors. Little data on the external validity of screens are available. DISCUSSION: Screens should be used with caution. A sequential procedure, in which subjects identified as being at risk during the first stage is followed by more specific diagnostic tests during the second stage, might overcome some of the limitations of the one-stage screening approach.