STATISTICAL METHODOLOGY--CASE CONTROL STUDIES OF PRIMARY HEPATOCELLULAR CANCER

统计方法--原发性肝细胞癌病例对照研究

• 批准号: 3816808
• 负责人: EDWARD D LUSTBADER
• 金额: --
• 依托单位: FOX CHASE CANCER CENTER
• 依托单位国家: 美国
• 资助国家: 美国
• 起止时间: 至
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/3816808
• 关键词: cancer risk; case history; disease vectors; epidemiology; hepatitis A; hepatitis B virus group; mathematical model; virus related neoplasm /cancer

We propose to investigate methods for inference and model criticism for relative risk regression models. We will first show how the likelihood of case-control and survival studies can be viewed as a realization of a multinomial random variable. This will suggest why there are difficulties in inference for general relative risk functions. These difficulties in inference for general reactive risk functions. These difficulties carry over to diagnostics where score based on the score to odds ratio estimation and then to investigate extensions to more general problems. It is anticipated that the cohort studies of humans and animals described in the other projects will be enhanced by having the most modern methods available for analysis of their data.

我们建议研究推理和模型批评的方法 相对风险回归模型。我们首先将展示以下可能性： 病例对照和生存研究可以被视为实现 多项式随机变量。这将说明为什么会出现困难 一般相对风险函数的推理。这些困难在 一般反应风险函数的推断。这些困难承载着 转到诊断，其中评分基于分数与优势比 估计，然后研究更一般问题的扩展。 预计人类和动物的队列研究将描述 其他项目将通过采用最现代的方法得到增强 可用于分析他们的数据。