Impacts of Geographic Distance on Peritoneal Dialysis Utilization: Refining Models of Treatment Selection.

OBJECTIVE: To examine the relationship between distance to dialysis provider and patient selection of dialysis modality, informed by the absolute distance from a patient's home and relative distance of alternative modalities. DATA SOURCES: U.S. Renal Data System. STUDY DESIGN: About 70,131 patients initiating chronic dialysis and 4,795 dialysis facilities in 2006. The primary outcome was patient utilization of peritoneal dialysis (PD). Independent variables included absolute distance between patients' home and the nearest hemodialysis (HD) facility, relative distance between patients' home and nearest PD versus nearest HD facilities, and their interaction. Logistic regression was used to model distance on PD use, controlling for patient and market characteristics. PRINCIPAL FINDINGS: Nine percent of incident dialysis patients used PD in 2006. There was a positive, nonlinear relationship between absolute distance to HD services and PD use (p < .0001), with the magnitude of the effect increasing at greater distances. In terms of relative distance, odds of PD use increased if a PD facility was closer or the same distance as the nearest HD facility (p = .006). Interaction of distance measures to dialysis facilities was not significant. CONCLUSIONS: Analyses of patient choice between alternative treatments should model distance to reflect all relevant dimensions of geographic access to treatment options.

Duke Scholars

Author Virginia Wang Population Health Sciences

Author Matthew Leonard Maciejewski Population Health Sciences

Author Cynthia Jan Coffman Biostatistics & Bioinformatics, Division of Biostatistics