Creation of a Risk-scoring Algorithm for Early Mortality Among Elderly ESRD Patie

Medical Technology And Practice Patterns, Bethesda MD

Investigators

Linked publications, trials & patents

Abstract

DESCRIPTION (provided by applicant): Creation of a Risk-scoring Algorithm for Early Mortality Among Elderly ESRD Patients Background. The rate of early deaths, those occurring within the three months following the diagnosis of end-stage renal disease (ESRD), are disproportionately high among patients who are 65 and older. Inexorable increases in the incidence of patients with both chronic kidney disease (CKD) and ESRD impel efforts to slow and reverse this urgent public health problem. This study aims to develop a risk-scoring system to aid in the clinical prediction of early death among elderly (>65 years) ESRD patients. Objective. We hypothesize that patients with high rates of early mortality soon after dialysis initiation have significant pre-dialysis medical conditions combined with inadequate renal disease management that does not prepare them appropriately for long- term dialysis. Our exploratory research project will identify factors associated with mortality within the first three months after starting permanent dialysis. We propose to examine the following associated with early mortality: 1) comorbid conditions found within one year prior to dialysis (e.g., cardiovascular events, acute renal failure, sepsis, hypertensive emergencies); 2) medical management prior to dialysis including involvement of a nephrology specialist, management of anemia, nutritional status, and vascular access preparation for dialysis that differentiates the 'emergent versus elective' (i.e., acute onset 'crash and burn' versus planned) transition to ESRD; and 3) use of health services in the latter stages of CKD. Results from these analyses will be used to develop a prognostic model and scoring algorithm for early mortality among dialysis patients. Methods. Using historical Medicare claims data, we propose to conduct a retrospective cohort study to develop a user friendly prognostic index. To identify the study population, we will construct a complete renal failure patient history by linking the pre-dialysis and post dialysis data to assess the predictors of early death. The proposed study is unique in linking the 2000-2009 ESRD data files to the 5% Chronic Kidney Disease (CKD) sample that has recently become available for research purposes. The total number of Medicare-eligible CKD patients from this period is approximately 831,000; of these, roughly 10% become ESRD-eligible and form the basis of our study population. Multivariate logistic regression models will b used to test each hypothesis and identify the significant risk factors for early mortality. Significance. We plan to identify potentially modifiable factors that are associated with early death among a population that transitions from CKD to ESRD requiring expensive, permanent and life-altering dialysis therapy. Given the high mortality rates during the first year after dialsis, accurate prediction of those destined for early death would be useful to patients and their families, providers, and society in making decisions about treatment.