Geographic variation in the termination of dialysis prior to death

This retrospective analysis of the 2004 United States Renal Data Service (USRDS) database examined how termination of dialysis prior to death varies geographically in the United States (US). Methods - Termination of dialysis prior to death was determined from the CMS-2746 - ESRD Death Notification form. The study population was restricted to patients with known date of death in 1995 to 2004, who had dialysis for 90+ days and lived in the US prior to death. Exclusion criteria included unknown age or gender, and age <20 or >110. Geographic location was defined by the patients' state. Unadjusted rates of termination were computed for the entire population, and then separately by age, race and gender. A mixed linear model with fixed effects was used to determine rates of termination adjusted for age, race and gender. US maps were prepared to demonstrate regional patterns. Results - Higher rates of termination were documented in Northern and Western states, and lower rates in Southern and Eastern states, ranging from 42% in Maine to 14% in Mississippi and 7% in the District of Columbia (DC). Whites terminated dialysis more frequently (26%) than blacks (13%) nationally and in all states and the DC that had adequate data for analysis. The lowest rates of termination were among patients in their 20s at the time of death (11%); the rates increased with age to 35% for those 90+ at the time of death. Conclusion - Termination of dialysis prior to death varies substantially by region in the US.