Dissertation/Thesis Abstract

Measurement Error Case Series Models for Assessing Infection-Cardiovascular Risk in Patients on Dialysis in the United States
by Mohammed, Sandra Marie, Ph.D., University of California, Davis, 2012, 235; 3544790
Abstract (Summary)

Infection and cardiovascular disease are leading causes of hospitalization and death in older patients on dialysis. It is of interest to determine if the incidence rate of cardiovascular events is increased for a period of time after an infection. The case series model is useful in studying the relationship between time-varying exposures, such as infections, and acute events, provides direct relative incidence estimation, controls for all measured and unmeasured time-invariant confounders and requires only cases. However, exact exposure times cannot be ascertained from available hospitalization data and can result in biased relative incidence estimates.

First, this work proposes a new method, the measurement error case series model, to handle data with imprecise times of exposure onset, i.e., exposure onset measurement error. The general nature of bias resulting from estimation that ignores measurement error is characterized and a bias-correction estimation method is proposed. Second, to assess inference, the validity and power of naive hypothesis testing is examined, which is based on applying the case series model to the imprecise data without correcting for the error. Third, a method for power/sample size determination is developed for planning case series studies where the exposure onset is measured with error.

The aforementioned methodologies are used to examine the infection-cardiovascular risk hypothesis in patients on dialysis in the United States. Hospitalization data from the United States Renal Data System, which captures nearly all (> 99%) patients with end-stage renal disease in the U.S. is utilized. The results suggest that the relative incidence estimate of cardiovascular events in the first 30 days following infections is substantially attenuated by 17-42% in the presence of infection onset measurement error.

Indexing (document details)
Advisor: Nguyen, Danh V.
Commitee: Dalrymple, Lorien S., Senturk, Damla
School: University of California, Davis
Department: Epidemiology
School Location: United States -- California
Source: DAI-B 74/03(E), Dissertation Abstracts International
Source Type: DISSERTATION
Subjects: Biostatistics, Statistics, Epidemiology
Keywords: Case series models, Exposure onset measurement errors, Longitudinal observational database, Non-homogeneous poisson process, Time-varying exposures, United states renal data system
Publication Number: 3544790
ISBN: 978-1-267-75971-9
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest