Minimally important difference for the Expanded Prostate Cancer Index Composite Short Form.

Objective To establish a score threshold that constitutes a clinically relevant change for each domain of the Expanded Prostate Cancer Index Composite (EPIC) Short Form (EPIC-26). Although its use in clinical practice and clinical trials has increased worldwide, the clinical interpretation of this 26-item disease-specific patient-reported quality of life questionnaire for men with localized prostate cancer would be facilitated by characterization of score thresholds for clinically relevant change (the minimally important differences [MIDs]). Methods We used distribution- and anchor-based approaches to establish the MID range for each EPIC-26 domain (urinary, sexual, bowel, and vitality/hormonal) based on a prospective multi-institutional cohort of 1201 men treated for prostate cancer between 2003 and 2006 and followed up for 3 years after treatment. For the anchor-based approach, we compared within-subject and between-subject score changes for each domain to an external "anchor" measure of overall cancer treatment satisfaction. Results We found the bowel and vitality/hormonal domains to have the lowest MID range (a 4-6 point change should be considered clinically relevant), whereas the sexual domain had the greatest MID values (10-12). Urinary incontinence appeared to have a greater MID range (6-9) than the urinary irritation/obstruction domain (5-7). Conclusion Using 2 independent approaches, we established the MIDs for each EPIC-26 domain. A definition of these MID values is essential for the researcher or clinician to understand when changes in symptom burden among prostate cancer survivors are clinically relevant.