The impact of different outcome measures on estimates of remission in a 3-year follow-up of eating disorders.

Objective: To examine the impact of using different outcome measures on estimates of remission rates in eating disorders (ED). Method: Adult ED patients (n = 334) included in a naturalistic, longitudinal project were assessed 3 years after intake. Remission was operationally defined using six different measures of outcome. Pairwise analyses of agreement between outcome measures were also calculated. Results: Remission rates for the entire group varied considerably from 24.3 to 77.8% depending on how remission was defined and who made the assessment (experts or patients). Outcome measures performed differently depending on diagnosis, and agreement between measures varied from none to very good. Conclusion: The impact of different definitions of outcome on remission estimates is considerable. There is a need to validate different estimates of remission by examining them in relation to multiple domains of outcome over time and by critically examining their relative merits both empirically and clinically. Copyright (C) 2010 John Wiley & Sons, Ltd and Eating Disorders Association.