OBJECTIVES: While patients with end-stage renal disease (ESRD) represent only 1 percent of Medicare beneficiaries, this population accounts for 8 percent of total Medicare spending. The disproportionate cost of treating ESRD has resulted in a high level of monitoring and policy focus. A number of public databases currently exist for study of the ESRD program in the US, but these databases are difficult to use and integrate for policy analysis of this program. Objective was to develop an integrated ESRD “data warehouse” with GIS capabilities from publicly-available dialysis patient and provider data and general population statistics, to enable a broadened scope of applications for policy analysis. METHODS: ESRD-related datasets were collected from Medicare sources and integrated into a relational database management system. Datasets included Renal and Hospital Cost Reports, Dialysis Facility Reports, the Dialysis Facility Compare database, and patient prevalence reports. Dialysis provider and patient addresses were geocoded and linked to spatial data files and general population data obtained from Census Bureau websites. Validation and outlier handling rules were created for most data and were applied either during the build/update process or at the point of analysis. Longitudinal datasets were created for trend analysis. Data are updated quarterly to keep the database current. RESULTS: The ESRD provider database contains over 6000 cost, utilization, quality, demographic and geographic variables, covering 100% of current Medicare ESRD providers (N=6288 facilities). Approximately 9% are hospital-based units and 91% are free-standing facilities. Depending on data source, the database covers between 6 and 14 calendar years (2001-14). Quality of the data varies, but for most applications, outlier and missing observations are less than 10%. CONCLUSIONS: Public datasets of ESRD-related information can be integrated with general population data and a GIS to support high-quality and cost-effective economic and clinical policy studies that would not otherwise be feasible.