Denis Mukwege, Lisa Peters, Christine Amisi, Alain Mukwege, Abigail R. Smith, Janis M. Miller

To derive a comprehensive system that allows a single score to define relative fistula severity. The present observational study included women with urogenital fistula treated at the Panzi Hospital, Democratic Republic of Congo, or its outreach clinics across the Democratic Republic of Congo between September 1, 2013, and December 31, 2014. Fistula severity was assessed by Goh and Waaldijk classifications and surgical success was ascertained. Logistic regression was used to select fistula characteristics predictive of surgical failure, and to preliminarily verify the newly derived Panzi score. Overall, 837 women were included in the analysis. Goh or Waaldijk fistula descriptors associated with a higher probability of poor surgical outcomes in the unadjusted bivariate analysis were circumferential defect (P=0.007), proximity to the external urethral orifice (P=0.001), and size (P=0.001). These fistula characteristics were used to construct the Panzi score, which varied from 3 (most severe) to 0 (minor fistula). For each increase above 0, the odds of surgical failure increase by a factor of 1.65 (P<0.001). The Panzi score of urogenital fistula provided a data-driven, simple, comprehensive, and parsimonious score. It could be used to report group data, to provide continuous level data for use in higher order statistics, and to resolve issues such as the cut-off point for referring women to hospital in accordance with fistula complexity. This article is protected by copyright. All rights reserved.