STATISTICAL METHODS IN AIDS RESEARCH

点击翻译按钮获取中文摘要

DESCRIPTION (The applicant's abstract) The overall objective of this grant is to develop and apply statistical methods in AIDS research. The focus is on the development of statistical methods for the analysis of longitudinal studies of HIV disease to improve scientific inferences from studies of HIV epidemiology, prevention and treatment. There are three specific aims of this proposal. The first specific aim is the development of methods for the analysis of data that are subject to informative censoring or more generally informative stochastic coarsening in longitudinal studies of HIV disease with time to event data. Right censored, interval censored and current status data are all examples of coarsened time to event data. This work will have applicability to studies of HIV prevention, treatment and epidemiology for estimation of HIV incidence rates, relative risks, vaccine and treatment efficacy. The work under specific aim 1 will be applied to a study of high risk HIV-negative individuals attending STD clinics in India, needle exchange evaluations in Baltimore, and a cohort of IV drug users in the U.S. (ALIVE). The second specific aim is the development of statistical methods for studies of maternal-infant transmission. This work will include regression models to estimate maternal-infant transmission rates that adequately account for loss to follow-up and differential infant mortality, as well as longitudinal data methods for tracking viral loads in maternal sera and breast milk and for correlating these loads with risks of maternal-infant transmission. The work under specific aim 2 will be applied to studies in Haiti and Zimbabwe. The third specific aim concerns the development of statistical methods for the analysis of recurrent events such as repeat hospitalizations and infections. The work under specific aim 3 will entail the development of one-sample and regression models for recurrent event data that address the issue of informative censoring for estimation of the occurrence rate of recurrent events over time and regression models. The methods will be applied to the ALIVE study.