Statistical Methods in HIV/AIDS Research

本项目旨在开发和应用新的统计方法，以改进艾滋病研究数据的科学推断，具体目标是：(1)建立一类通用的回归模型，研究艾滋病治疗和预防效果试验中的剩余时间。开发的方法将用于在剩余时间内确定早期抗逆转录病毒治疗与延迟抗逆转录病毒治疗对患者到艾滋病定义疾病或死亡的时间是否重要，并进一步探索在开始抗逆转录病毒治疗后控制艾滋病进展或死亡的重要临床因素；(2)开发统计方法，用于重复事件和标记过程数据，这些数据在信息故障事件之前表现出终端行为。开发的方法将用于检查治疗方案与复发事件（如机会性感染）的完整病史之间的相关性，同时考虑到临床医生在常规护理过程中可用的变量；(3)发展统计方法，从艾滋病随机预防试验的复杂纵向数据中推断因果治疗效果。

DESCRIPTION (provided by applicant): Statistical Methods in HIV/AIDS Research Project Summary The objective of this methods proposal is to develop and apply new statistical methods to improve scientific inferences for data collected in HIV/AIDS research, specifically aiming to: (1) develop a general class of regression models to study residual time in HIV treatment and prevention efficacy trials. The developed methods will be applied to determine in residual time whether early versus delayed ART initiation is important in patient's time to AIDS-defining illness or death, and to further explore the clinical factors which are important in governing the progression to AIDS or death after ART initiation; (2) develop statistical methods for recurrent event and marker process data that exhibit a terminal behavior before an informative failure event. The developed methods will be applied to examine the correlation between treatment regimens and the complete history of recurrent events such as opportunistic infections, taking into account variables available to clinicians in the course of routine care; (3) develop statistical methodologies to infer causal treatment effect from complex longitudinal data in randomized HIV prevention trials. PUBLIC HEALTH RELEVANCE: Statistical Methods in HIV/AIDS Research Project Narrative This proposal aims to develop statistical methods in HIV/AIDS research. Specifically, these methods are targeted to regression analysis of censored residual life, analysis of recurrent markers in the presence of terminal events, and informative censoring and causal inferences of treatment efficacy in complex HIV/AIDS clinical trials. Upon successful completion of the project, the proposed methods shall provide a collection of new statistical tools for resource planning, clinical consulting and efficacy assessment in HIV/AIDS research.