Abstract
There is substantial geographic variability in both incident and prevalent arteriovenous fistula (AVF) use among patients with ESRD. This study examined the degree to which these variations associate with poverty in the county of a patient's treatment center. We performed a cross-sectional study including 28,135 patients treated by 1127 hemodialysis centers in five ESRD networks (16 states) between June 1, 2005 and May 31, 2006. We used the 2000 U.S. Census to categorize county-level poverty and ascertained incident AVF use from the Medicare CMS 2728 form. We calculated the 30-month slope of change in AVF prevalence from monthly facility reports collected between 2003 and 2005. More than 33% of treatment centers were located in high-poverty counties. County poverty inversely associated with incident AVF use (P for trend = 0.001). In contrast, substantial increases in prevalent AVF rates from 30.9 to 38.6% (P < 0.001) among treatment centers did not associate with county poverty (P = 0.9519). In conclusion, the concentration of poverty in the county where a treatment is located associates with incident AVF use by patients with ESRD but not with subsequent improvement in AVF use among prevalent patients. These results suggest that the Medicare ESRD program may mitigate poverty effects on AVF use.