Biostatistics

生物统计核心B将为所有4个莫菲特皮肤孢子项目、其他2个核心以及发展研究和职业发展计划提供协作分析支持。每个项目都有一名生物统计师，他们的选择是为了最大限度地将他们的应用研究重点与项目之间的联系最大化。将有4项临床试验，包括项目2和3中的第一阶段试验，这些试验将使用改进的JI设计，以提供建立MTD的更丰富的体验。这一设计得到了广泛的讨论，并包括核心B统计学家的一些修改，将受到特别的持续关注，以适应一种新的分析方法。从分析的角度来看，项目2和项目4都非常复杂，将获得最高比例的资金。项目2涉及广泛的蛋白质组学分析，并使用了部分由项目2统计学家开发的多水平贝叶斯模型。项目4是一项病例队列研究，涉及1500名参与者和3000人年的跟踪调查，其中包括许多需要检查的人口统计和生活方式因素。与A和C核心一起，该核心将在孢子数据库的开发中发挥至关重要的作用，特别是在临床试验和病例队列研究方面。生物统计学核心B将参与所有拟议的发展研究和职业发展计划的发展，并将为那些得到资助的项目提供分析支持。在学院生物统计师的指导下工作的是两名员工统计员，他们将处理所需的直接统计编程的相当大份额。

Biostatistics Core B will provide collaborative analytic support to all 4 Moffitt Skin SPORE projects, the other 2 Cores, and the Developmental Research and Career Development Programs. Each project has a faculty biostatistician, selected to maximize the link between their applied research focus and the project. There, will be 4 clinical trials, including phase 1 trials in Projects 2 and 3 that will use a modified Ji design, chosen to provide a richer experience in establishing an MTD. This design, which received extensive discussion and included some modification by the Core B statisticians, will receive special ongoing attention as befits a novel analytical approach. Projects 2 and 4 are both very involved from an analytical perspective and will receive the highest proportions of funding. Project 2 involves extensive proteomics analyses and uses a multi-level Bayesian model developed in part by the project 2 statistician. Project 4 is a case-cohort study involving 1500 participants and 3,000 person-years of follow-up, and includes numerous demographic and lifestyle factors to be examined. In conjunction with Cores A and C, this core will play a vital role in SPORE database development, especially with the clinical trials and case-cohort studies. Biostatistics Core B will be involved in the development of all proposed Developmental Research and Career Development Programs, and will provide analytic support to those that are funded. Working under the direction of faculty biostatisticians are two staff statisticians, who will handle a significant share of the direct statistical programming needed.