Geography matters: Relationships among urban residential segregation, dialysis facilities, and patient outcomes

Background: End-stage renal disease disproportionately affects black Americans. However, the impact of residential segregation by race-a prominent feature of many U.S. cities-on outcomes of patients receiving dialysis and on facility performance has not been evaluated. Objective: To examine the relationship among racial composition of ZIP codes in metropolitan areas, outcomes of patients receiving dialysis, and characteristics of dialysis facilities. Design: Retrospective cohort study of patients receiving dialysis and cross-sectional study of dialysis facilities. Setting: U.S. metropolitan ZIP codes with differing percentages of black residents. Patients: Black and non-Hispanic white patients who initiated long-term dialysis between 1 January 1995 and 31 December 2002 (n = 399 424) and dialysis facilities in operation in December 2004 (n = 3244). Measurements: Mortality and time to transplantation among patients receiving dialysis, and performance of dialysis facilities on the basis of quality indicators (anemia management, dialysis adequacy, and facility-level mortality rates). Results: Most black patients (50.3%) but few white patients (5%) lived in the 3% (n = 769) of ZIP codes in which most residents were black. in analyses adjusted for patient and ZIP code characteristics, mortality rates were higher among white patients but not among black patients living in areas with a higher percentage of black residents (adjusted hazard ratio for ZIP codes with 75% black residents vs. < 10% black residents, 1.14 [95% Cl, 1.07 to 1.21] for white patients and 1.02 [Cl, 0.99 to 1.06] for black patients). Time to transplantation was longer among both black and white patients (adjusted hazard ratio for ZIP codes with >= 75% black residents vs. < 10% black residents, 0.84 [Cl, 0.78 to 0.92] and 0.63 [Cl, 0.57 to 0.71] for black patients and white patients, respectively). Dialysis facilities located in areas with a higher percentage of black residents were more likely to have higher-than-expected mortality rates and were less likely to meet performance targets. Limitations: Patient-level analyses were restricted to black and non-Hispanic white patients. Patient-level and facility-level analyses focused only on the percentage of black residents in each ZIP code. Conclusions: The racial composition of urban residential areas is associated with time to transplantation and dialysis facility performance on standard quality measures. Closer scrutiny of care provided to patients receiving dialysis who live in predominantly black residential areas and to dialysis facilities operating in these areas may be warranted.