Recovery from anorexia nervosa

The aim of this chapter is to critically analyze the difficulties still inherent in studying recovery from anorexia nervosa (AN). A comprehensive literature review demonstrates that there are at least ten definitions of recovery from AN, differentiated according to several criteria: definition of normal weight; inclusion of only behavioral vs. both behavioral and psychological recovery; inclusion of comorbidity and overall psychosocial functioning in addition to eating-related recovery; and the minimal time required to define recovery. These discrepancies likely account for the wide variation in the range of patients defined as recovered from AN in different outcome studies (0-92%). A great variability also exists with respect to the factors predicting recovery. Another limitation to the research on factors that predict recovery in AN is that most outcome studies have used univariate analyses for this purpose, enabling the assessment of only separate independent predictors. By contrast, multivariate quantitative models allow for the measurement of the relative contribution of each variable for recovery, as well as for the assessment of the cumulative predictive power of all variables included. Unfortunately, only a few quantitative studies exist. Still, an overall worse prognosis has been found repeatedly to be associated with several factors: body image disturbances, compulsive exercising, comorbid anxiety disorders (particularly obsessive compulsive disorder), personality disorders and non-clinical personality traits, poor social skills, longer duration of illness prior to treatment, and shorter follow-up. In the second part of the chapter we aim to provide the reader with several potential options to standardize the definition of recovery and the design of outcome studies in AN.