Objective: To develop an empirically derived definition of mixed urinary incontinence (MUI) for use in incontinence outcomes research. Methods: Participants in a randomized trial comparing the fascial sling and. Burch colposuspension were assessed using standardized measures including the Medical, Epidemiologic and Social Aspects of Aging (MESA), UI questionnaire, the Urogenital Distress Inventory (UDI), 3-day urinary diary and urodynamic studies (UDS). Participants were required to have stress incontinence with a MESA stress subscale score > MESA urge subscale score. Several definitions of MUI were considered. Logistic and linear regression analysis methods were used to predict clinical outcomes based on the different MUI definitions. Analyses were carried out using SAS (SAS Institute, Inc., Cary, NC, Version 9.1). Statistical significance was defined at P-value <0.05. Results: In 655 participants, the proportion of women with MUI varied from 8.3% to 93.3% depending on the MUI definition All definitions were associated with severity as measured by the frequency of incontinence episodes at baseline; however little of the variability was explained by any single definition. No strict cut-off value for these baseline measures was identified to predict clinical outcomes. Conclusions: These MUI definitions do not adequately categorize clinically relevant UI subgroups. For research reporting, MUI subcomponents of stress and urge UI should be described separately rather than as a single dimension. © 2009 Wiley-Liss, Inc.