It is estimated that more than 230,000 Americans have end-stage renal disease (ESRD). The morbidity and mortality associated with ESRD are major public health problems, and must be addressed. A traditional focus of the nephrology literature has been on mortality-based comparisons between various subgroups of the ESRD patient population. More recent trends suggest that the study of pre-mortality outcomes may potentially be more useful. Analyses to date have primarily been descriptive (e.g. tabular), and compare characterizations of morbidity experienced across subgroups of the ESRD patient population. The use of morbidity as an outcome is clearly sensible; it is reasonable to expect that a therapy resulting in reduced levels of morbidity should also decrease mortality, and has the added benefit of improving quality of life, something directly relevant and important to ESRD patients. There is a need to develop and employ more sophisticated methodology for analyzing the morbidity and mortality patterns within the current ESRD patient population. The results of such analyses may be used to recommend treatment sequences for various ESRD subpopulations, and perhaps identify comorbid factors that new treatments should be designed to address; both have the potential to reduce morbidity and mortality associated with ESRD. Therefore, this proposal has two major objectives. The first involves an in-depth investigation into the use of hospitalization patterns as an alternative outcome measure to mortality. The decision to focus on hospitalization is due to its strong correlation with mortality (thus serving as a possible predictive marker for mortality), and its potential as a longitudinally informative description of patient morbidity. The second objective, a necessary step in the direction towards achieving the first, involves the development of more sophisticated methods for analyzing both outcomes. Specifically, the intent is to develop methodology that is able to (i) effectively evaluate hospitalization as a marker for mortality; (ii) model morbidity processes using longitudinal outcomes like hospitalization; and, (iii) improve upon current analyses using mortality as the outcome.
| Strawderman, R L; Parzen, M I; Wells, M T (1997) Accurate confidence limits for quantiles under random censoring. Biometrics 53:1399-415 |
| Strawderman, R L (1997) An asymptotic analysis of the logrank test. Lifetime Data Anal 3:225-49 |
| Strawderman, R L; Levine, G; Hirth, R A et al. (1996) Using USRDS generated hospitalization tables to compare local dialysis patient hospitalization rates to national rates. Kidney Int 50:571-8 |
