Statistical Methods and Theory for Clinical Trials in the Era of Patient-Oriented Research and Personalized Medicine

我的研究计划的主要目标是在临床试验的设计和分析中为统计方法奠定坚实的数学基础，以评估预防和治疗疾病的新药或程序，如果没有方法，则开发新的统计方法，并对其数学特性进行全面调查。近年来，以患者为中心的个性化药物的研究和开发是临床试验的两个重点领域。以患者为中心的研究旨在让患者参与研究的设计、实施和解释，包括通过回答特定问卷上的问题来评估患者报告的治疗结果的效果，而个性化医学旨在根据每个患者的个体特征发展医疗。设计和分析这两个研究重点的临床试验需要新的统计方法和理论。在上一个授予周期中，我们在这一统计研究领域做出了重大贡献。我们计划继续这项研究计划，重点研究我们最近在实践中遇到的以下四个问题：(1)开发混合密度比模型和相关推理程序，用于分析限制在封闭区间内的纵向患者报告结果；(2)开发统计方法，根据基于连续和/或多个生物标记物的患者报告的结果，识别可能受益于或受到新治疗伤害的患者亚群；(3)对重采样程序进行理论评估，以评估通过套索和相关程序选择的生物标记物对事件结果的影响；(4)基于治疗效果的非参数测量，开发具有多个生物标志物和事件发生前时间的亚组分析的统计程序。预计这项研究将为在临床试验和其他相关领域工作的应用统计学家提供一些新的统计工具，并为这些新方法奠定坚实的数学基础，从而更有效和可靠地确定预防和治疗疾病的新战略，从而减少这些疾病的发病率和死亡率。预计它还将促进概率和统计数学理论的新发展。本项目所发展的理论和方法可应用于其他类型的临床研究，也可应用于其他自然科学和工程领域，如具有检测极限的物理化学试验和可靠性工程。将通过参与本提案所列项目，培训各级高素质人员。

The primary objectives of my research program are to lay solid mathematical foundations for statistical methods in the design and analysis of clinical trials for the evaluation of new drugs or procedures in the prevention and treatment of the diseases and, if methods are not available, to develop new statistical methods with a through investigation of their mathematical properties.  In recent years, patient-orientated research and development of personalized medicines are two focus areas in the clinical trials. Patient-orientated research intends to involve patients in the design, conduct, and interpretation of a study including the assessment on the effects of treatments by outcomes reported by patients through answers to questions on specific questionnaires, while personalized medicine aims the development of medical treatment tailored to the individual characteristics of each patient. New statistical methods and theory are required for the design and analysis of clinical trials with these two research focuses.  In the last granting cycle, among other achievements, we have made major contributions in this area of statistical research. We plan to continue this research program with focus on the study of the following four problems we have encountered recently in our practices: (1) Development of a mixture density ratio model and associated inference procedures for the analysis of longitudinal patient reported outcomes which are restricted to a closed interval; (2) Development of statistical methods which can be used to identify subgroups of patients who may benefit from or be harmed by a new treatment with respect to outcomes reported by patients based on continuous and/or multiple biomarkers; (3) Theoretical evaluations of resampling procedures for the assessment of the effects on time to an event outcome for biomarkers which are selected by LASSO and related procedures; (4) Development of statistical procedures for subgroup analyses with multiple biomarkers and time to an event outcomes based on a nonparametric measurement of treatment effects. It is expected that this research will generate some new statistical tools for applied statisticians working in clinical trials and other related areas and also lay solid mathematical foundations for these new methods, which will result in more efficient and reliable identification of new strategies for the prevention and treatment of diseases and, therefore, reduction of morbidity and mortality from these diseases. It is also expected it will stimulate new development in the mathematical theory of probability and statistics. The theory and methods developed from this project may be applied to other types of clinical studies and also other fields of natural science and engineering such as physical and chemical testing with limits of detection and reliability engineering. All levels of Highly Qualified Personnel will be trained through the participation in the projects listed in this proposal.