Over 215,000 patients with end-stage renal failure in the U.S. are currently receiving renal replacement therapy, a treatment which in 1990 was estimated to cost 7.26 billion dollars. Despite this great expenditure, mortality rates remain high. The gross mortality rate in 1989-1991 was 24 patients per 100 patient years. This mortality risk has been shown to be higher among whites compared to blacks and among males compared to females. These differences in mortality has been attributed to specific causes of death. Potential explanations include differences in comorbidity at initiation of ESRD, differences in treatment, physiological differences or social/cultural differences between racial and gender subgroups. We plan to explore the hypotheses that 1) these differences are in part due to a greater degree of comorbidity among whites and males compared respectively to blacks and females at the initiation of ESRD and/or 2) these differences are due to differences in treatment by race and gender. Previous studies have identified predictors of all-cause mortality in the ESRD population. We hypothesize that predictors of all cause and cause specific mortality differ by subgroup. We therefore propose to describe differences in comorbidity by race and gender and determine their effect on all cause and cause specific mortality. We also propose to identify predictors of all cause and cause- specific mortality within gender and racial subgroups. To achieve these objectives we plan to undertake a retrospective analysis of previously collected data from the USRDS Case Mix Severity and Case Mix Adequacy studies. The Case Mix severity dataset includes demographic comorbidity and outcome data on a random sample of over 4800 patients starting dialysis for ESRD during 1986 and 1987. The Case Mix Adequacy Study includes similar data with the addition of data on treatment on a random sample of over 7000 hemodialysis patients prevalent or incident in 1990. In our analyses, the dependent variables studied will be death, time to death and cause of death. The independent variables will include age, gender, race, dialysis modality, comorbidity and treatment factors including dose of dialysis. Comorbidity will be measured with 9 individual comorbid conditions which will be collapsed from original variables in the dataset. In addition a comorbidity index, a measure of overall comorbidity, which is currently under development will be utilized. Dose of dialysis will be calculated utilizing a number of treatment parameters included in the dataset. Cox Proportional Hazards Modelling will the major statistical analytic technique that will be utilized. Logistic regression and ordinary least squares regression will be utilized as appropriate.
| Bloembergen, W E; Hakim, R M; Stannard, D C et al. (1999) Relationship of dialysis membrane and cause-specific mortality. Am J Kidney Dis 33:1-10 |
