As the mean age of incident dialysis patients has increased to 60 years and prevalence of comorbid conditions such as diabetes has increased, increasing comorbidity may profoundly impact on morbidity, mortality and costs in the dialysis population. A valid tool that quantifies comorbidity may help to identify high-risk patients with expensive resource utilization, help to flame public health strategies and define optimal dialysis therapeutic interventions targeted towards this high-risk group. To be valid in the dialysis population, the comorbidity tool should be simple, readily verifiable, include certain specific factors that influence outcomes in dialysis patients such as duration of end-stage renal disease and include objective, quantifiable measures of disease severity. None of the currently used indices such as the Charlson Comorbidity Index and the Index of Coexistent Diseases meet all of these criteria. Therefore, the specific aims of this study are to develop and validate in a large sample of dialysis population an index of comorbidity that is simple, specific for dialysis patients and accounts for disease severity. Prior studies showed patients with high comorbidity have higher peritoneal dialysis technique failure. This study will also examine whether the increased mortality observed after peritoneal dialysis technique failure is due to high comorbidity of these patients or the technique failure itself causes higher mortality. This study will use the existing data in the United States Renal Data System (USRDS) Dialysis Morbidity and Mortality Studies (DMMS). The comorbidity tool will be developed from a subset of the 6300 prevalent patients in the DMMS III study. By Cox proportional hazards, the relative risk for death, for each of the comorbid conditions will be determined and scores will be given based on the relative risk. This scoring system will be validated in the following sub-populations: prevalent hemodialysis patients with Medicare as the primary or sole payer in the DMMS-IV study, prevalent hemodialysis patients with non-Medicare as the primary or sole payer in the DMMS-IV study, incident hemo and peritoneal dialysis started on dialysis in 1996 or 1997 in the DMMS wave II study irrespective of the insurance status. The outcomes of interest will be hospital days and Medicare hospital costs (by ANOVA) and death (by Cox proportional hazards). Further data from DMMS II will be used to examine the risk of death of incident peritoneal dialysis patients who failed the technique with incident hemodialysis patients. As comorbidity predicts peritoneal dialysis technique failure, the above analysis will help in determining the optimal dialysis modality based on comorbidity at the initiation of dialysis.
