STATISTICAL ISSUES IN AIDS RESEARCH

描述(摘自摘要)：在这个项目中，研究人员 处理试验数据统计分析中的重要问题 治疗或预防艾滋病毒感染的干预措施，继续 在本批款期内开展的临床治疗方法方面的工作 艾滋病治疗试验，并将工作范围扩大到 现场试验中出现的设计分析问题的探讨 预防艾滋病毒感染的疫苗。艾滋病危机导致了 迫切需要迅速获得有关疗效和安全性的数据 治疗干预措施。设计高效的临床试验是 最可靠的信息来源，其有效评估 治疗效果可以根据。为了提高效率和 试验的可靠性，首席调查员建议继续 致力于探索从使用代理端点信息的方法 生物标志物，如CD4和RNA聚合酶链式反应图谱。他期待着 探索确定长期治疗效果的方法 由于最终结果，随机临床试验已提前终止 短期影响的证据。工作将继续进行， 开发需要处理的重要方法 适当地，结果数据的类型是多变量的或 受制于复杂的审查机制，而且经常出现 在艾滋病队列研究以及艾滋病毒临床试验中。具体来说， 将开发非参数和半参数方法进行分析 将开发多变量故障时间数据和统计方法 用于区间截尾、双截尾和截尾的分析 故障时间数据。这些方法将使用以下数据进行评估 艾滋病毒/艾滋病临床试验和队列研究。艾滋病毒的其他特征 疫苗试验也将得到解决，包括 疫苗试验参与者对疫苗的依赖性和异质性 疫苗效力估计量的分布特性。新的 将开发和评估统计方法，用于分析 影响艾滋病毒试验参与者的病毒特征依次为 作出有关豁免范围的推论 疫苗。

DESCRIPTION (adapted from Abstract): In this project, the researchers address important issues in the statistical analysis of data from trials of interventions for treatment or prevention of HIV infection, continuing work conducted during the current grant period on methods of clinical trials of AIDS therapeutics, and extending the work scope to the exploration of design and analysis issues arising in field trials of vaccines for prevention of HIV infection. The AIDS crisis has led to the urgent need for rapidly obtaining data about the efficacy and safety of therapeutic interventions. Efficiently designed clinical trials are the most reliable source of information on which valid assessments of treatment efficacy can be based. To improve the efficiency and reliability of trials, the Principal Investigator proposes to continue work on exploring methods to use surrogate endpoint information from biological markers, such as CD4 and RNA PCR profiles. He expects to explore approaches for determining long-term treatment effects when randomized clinical trials have been terminated early due to definitive evidence of short-term effects. Work will be continued on the development of important methods which are required to handle appropriately the type of outcome data which are multivariate or are subject to complicated censoring mechanisms and which arise frequently in AIDS cohort studies as well as in HIV clinical trials. Specifically, non- and semi-parametric methods will be developed for analyzing multivariate failure time data and statistical methods will be developed for the analysis of interval censored, doubly censored, and truncated failure time data. These methods will be assessed using data from HIV/AIDS clinical trials and cohort studies. Other features of HIV vaccine trials will also be addressed, including the effects of dependence and heterogeneity among vaccine trials participants on distributional properties of estimators of vaccine efficacy. New statistical methods will be developed and evaluated for the analysis of characteristics of viruses that impact HIV trial participants in order to make inferences regarding breadth of immunity conferred by the vaccine.