Initial Validation of AUGS-PERFORM: Construct Validity and Test-Retest Reliability

Abstract

Importance The American Urogynecologic Society’s Prolapse pERFORmance Measure (AUGS-PERFORM) patient-reported outcome measure contains 11 items designed to assess symptoms relevant for assessing the quality of treatment for pelvic organ prolapse. Objective The aim of the study was to test AUGS-PERFORM’s construct validity and test-retest reliability. Study Design For this prospective validation study, we recruited English-speaking adult participants, at a single academic institution, seeking care for pelvic organ prolapse. Participants completed AUGS-PERFORM, the Pelvic Floor Distress Inventory, and several Patient-Reported Outcomes Measurement Information System short forms at baseline. We compared the 11 AUGS-PERFORM items against items testing the same concepts on other questionnaires using percent agreement, kappa statistics, and linear regression to determine construct validity. Two weeks later and before any pelvic organ prolapse therapy, participants completed AUGS-PERFORM a second time. Test-retest reliability was assessed using intraclass correlation coefficients. Results We enrolled 148 participants between 27 and 86 years of age: 81% self-identified as White, 56% were sexually active, and 84% elected surgery for treatment. The AUGS-PERFORM items assessing bulge presence and bother had a high percent agreement with the Pelvic Floor Distress Inventory item #3 (83.5 and 70%, respectively). The percent agreement ranged from 69% to 75% for items assessing urinary and defecatory symptoms and from 49% to 56% for pain-related questions. Sexual function items had a strong negative correlation (expected direction based on scoring) with validated sexual function questionnaires. The intraclass correlation coefficient was estimated to be 0.86, indicating excellent test-retest reliability. Conclusions The AUGS-PERFORM demonstrated good construct validity for prolapse, urinary incontinence, defecatory dysfunction, and sexual function questions, and moderate construct validity for pain-related questions. There was excellent test-retest reliability.