More than a third of middle-aged or older women suffer from urinary incontinence, but less than half undergo evaluation or treatment for this burdensome condition. With national organizations now including an assessment of incontinence as a quality performance measure, providers and health care organizations have a growing incentive to identify and engage these women who are undiagnosed and untreated. We sought to identify clinical and sociodemographic determinants of patient-provider discussion and treatment of incontinence among ethnically diverse, community-dwelling women. We conducted an observational cohort study from 2003 through 2012 of 969 women aged 40 years and older enrolled in a Northern California integrated health care delivery system who reported at least weekly incontinence. Clinical severity, type, treatment, and discussion of incontinence were assessed by structured questionnaires. Multivariable regression evaluated predictors of discussion and treatment. Mean age of the 969 participants was 59.9 (±9.7) years, and 55% were racial/ethnic minorities (171 black, 233 Latina, 133 Asian or Native American). Fifty-five percent reported discussing their incontinence with a health care provider, 36% within 1 year of symptom onset, and with only 3% indicating that their provider initiated the discussion. More than half (52%) reported being at least moderately bothered by their incontinence. Of these women, 324 (65%) discussed their incontinence with a clinician, with 200 (40%) doing so within 1 year of symptom onset. In a multivariable analysis, women were less likely to have discussed their incontinence if they had a household income < $30,000/y vs ? $120,000/y (adjusted odds ratio [AOR], 0.49, 95% confidence interval [CI], 0.28-0.86) or were diabetic (AOR, 0.71, 95% CI, 0.51-0.99). They were more likely to have discussed incontinence if they had clinically severe incontinence (AOR, 3.09, 95% CI, 1.89-5.07), depression (AOR, 1.71, 95% CI, 1.20-2.44), pelvic organ prolapse (AOR, 1.98, 95% CI, 1.13-3.46), or arthritis (AOR, 1.44, 95% CI, 1.06-1.95). Among the subset of women reporting at least moderate subjective bother from incontinence, black race (AOR, 0.45, 95% CI, 0.25-0.81, vs white race) and income < $30,000/y (AOR, 0.37, 95% CI, 0.17-0.81, vs ? $120,000/y) were associated with a reduced likelihood of discussing incontinence. Those with clinically severe incontinence (AOR, 2.93, 95% CI, 1.53-5.61, vs low to moderate incontinence by the Sandvik scale) were more likely to discuss it with a clinician. Even in an integrated health care system, lower income was associated with decreased rates of patient-provider discussion of incontinence among women with at least weekly incontinence. Despite being at increased risk of incontinence, diabetic women were also less likely to have discussed incontinence or received care. Findings provide support for systematic screening of women to overcome barriers to evaluation and treatment.