Race and end-stage renal disease in the United States Medicare population: The disparity persists

Charles R V Tomson, Robert N. Foley, Qi Li, David T. Gilbertson, Jay L. Xue, Allan J. Collins

Research output: Contribution to journalArticlepeer-review

8 Scopus citations


Aim: A marked preponderance of end-stage renal disease among African Americans was described more than two decades ago. The objective of this study was to determine whether racial disparities in end-stage renal disease in the United States have changed over time. Methods: The authors compared renal replacement therapy rates in five biennial cohorts (1993-1994, 1995-1996, 1997-1998, 1999-2000, 2001-2002; n = 6 315 283), using annual random samples of 5% of the US Medicare population and the United States Renal Data System registry. Results: The proportion of African American subjects rose from 8.8% in the first cohort to 9.4% in the last. Renal replacement therapy rates (per 1000 patient-years) among white Americans in successive cohorts were 0.84, 0.96, 1.08, 1.16 and 1.20, compared with 2.98, 3.24, 3.65, 3.80 and 3.57 among African Americans (P < 0.0001 for race comparison within each biennial cohort). Corresponding hazards ratios, adjusted for demographic characteristics and comorbid conditions, were 2.01 (95% confidence interval 1.82-2.33), 1.96 (1.78-2.17), 2.00 (1.81-2.20), 2.01 (1.83-2.21) and 1.86 (1.69-2.04), suggesting the absence of meaningful reduction in racial disparities in renal replacement therapy rates over time. Conclusion: Disparities in renal replacement therapy rates between white and African American Medicare beneficiaries have persisted over time.

Original languageEnglish (US)
Pages (from-to)651-656
Number of pages6
Issue number7
StatePublished - Oct 2008


  • African Americans
  • End-stage renal disease
  • Medicare
  • Race
  • Renal replacement therapy


Dive into the research topics of 'Race and end-stage renal disease in the United States Medicare population: The disparity persists'. Together they form a unique fingerprint.

Cite this