Mortality among hemodialysis patients in Europe, Japan, and the United States: Case-mix effects

Abstract

The Dialysis Outcomes and Practice Patterns Study is well suited to identify case-mix effects, given its extensive data set. The data set was used to examine the influence of case-mix variables on mortality and the extent to which these variables account for differences in mortality across regions, as well as the prevalence and incidence of hepatitis B and hepatitis C. Demographic and comorbid disease features were determined for 8,615 patients internationally; mortality was recorded for this cohort, plus replacement patients (total n = 16,720), from 1996 to 2002. Mortality was associated with increasing age, nonblack race, coronary artery disease, congestive heart failure, other cardiac disease, diabetes mellitus, peripheral vascular disease, cerebrovascular disease, absence of hypertension, lung disease, cancer, human immunodeficiency virus infection, gastrointestinal bleeding, neurologic disease, psychiatric disease, cellulitis/gangrene, hepatitis C, and smoking. US patients were slightly older than those in Europe or Japan and had the highest prevalence of diabetes, coronary artery disease, congestive heart failure, peripheral vascular disease, and cerebrovascular disease. Upon adjusting for case-mix to assess mortality across facilities, it was found that regional differences in mortality (highest in the United States and lowest in Japan) and differences across facilities within nations remain after such corrections. It is likely that practice patterns account for some of this variation. Prevalence of hepatitis B virus (HBV) across facilities increased as the number of dialyzing patients per facility increased; risk of HBV seroconversion decreased among facilities using protocols for treatment of patients with HBV infection. Greater employment of staff with at least 2 years of formal nursing training was associated with lower prevalence of hepatitis C virus infection and lower seroconversion risk.