Identifying duration criteria for eating-disorder remission and recovery through intensive modeling of longitudinal data.

Objective: Outcome states, such as remission and recovery, include specific duration criteria for which individuals must be asymptomatic. Ideally, duration criteria provide predictive validity to outcome states by reducing symptom-return risk. However, available research is insufficient for deriving specific recommendations for remission or recovery duration criteria for eating disorders.

Method: We intensively modeled the relation between duration criteria length and rates of remission, recovery, and subsequent symptom return in longitudinal data from a treatment-seeking sample of women with anorexia nervosa (AN) and bulimia nervosa (BN). We hypothesized that the length of the duration criterion would be inversely associated with both rates of remission and recovery and with subsequent rates of symptom return.

Results: Generalized estimating equations supported our hypotheses for all investigated eating-disorder features except for symptom return when using the Psychiatric Status Rating for AN.

Discussion: We recommend that 6 months be used for remission definitions applied to binge eating, purging, and BN symptom composite measures, whereas no duration criteria be used for low weight and AN symptom composites. We further recommend that 6 months be used for recovery definitions applied to BN symptom composites and AN symptom composites, whereas 18 months be used for individual symptoms of binge eating, purging, and low weight. The adoption of these duration criteria into comprehensive definitions of remission and recovery will increase their predictive validity, which in turn, maximizes their utility.