Statistical Methods for Estimation of Benefits & Harms of Repeat Cancer Screening

描述（由申请人提供）：超过40%的美国人在其一生中会患上癌症，大约五分之一的人会死于癌症。然而，筛查可以降低癌症死亡率并降低某些癌症的发病率。确定有效的筛查方案--谁应该接受癌症筛查，多久一次，用什么方法--是公共卫生的关键。用于做出这些决定的大部分数据与个人筛选测试的性能有关。相对较少的研究集中在比较替代的筛查方案，但这些信息是需要提供证据的筛查指南。本研究的总体目标是开发统计方法，用于表征包括方案的多轮筛查后的危害（假阳性检测结果和遗漏的癌症）和益处（筛查检测到的癌症），以促进指南制定和决策制定。我们的工作将涉及三个具体目标：（1）开发新的统计方法，用于同时估计重复癌症筛查的危害和益处的累积风险;（2）开发新的统计方法，用于估计在筛查方案过程中经历的假阳性测试结果的预期数量;（3）使用我们新开发的统计方法来分析基于风险的筛查乳房X线摄影方案，使用来自乳腺癌监测联盟（BCSC）的15年数据。由于现有的方法无法解释观察计划的重要特征，因此以前不可能对针对女性个体乳腺癌风险水平的重复乳腺癌筛查的危害和益处进行无偏估计。根据目标3，我们将使用根据目标1和2开发的统计方法分析BCSC的数据，并将比较针对女性乳腺癌风险特征的替代乳腺癌筛查方案。通过提供关于基于风险的乳腺癌筛查方案的潜在危害和益处的信息，这项工作将为决策者提供信息，并促进患者和提供者之间的沟通。这将有助于个人决策者和循证指南的制定。

DESCRIPTION (provided by applicant): More than 40% of Americans will develop cancer in their lifetime and approximately 1 in 5 will die of cancer. However, screening can reduce cancer mortality and decrease incidence for some cancers. Identifying effective screening regimens- who should be screened for cancer, how often, and by what method- is key to public health. Most of the data used to make these decisions relate to the performance of individual screening tests. Relatively little research has focused on comparing alternative screening regimens, yet this information is needed to provide evidence for screening guidelines. The overall goal of this research is to develop statistical methods for characterizing harms (false-positive test results and missed cancers) and benefits (screen- detected cancers) after the multiple rounds of screening comprising a regimen in order to facilitate guideline setting and decision making. Our work will address three specific aims: (1) To develop new statistical methods for simultaneously estimating the cumulative risk of harms and benefits of repeat cancer screening; (2) To develop new statistical methods to estimate the expected number of false-positive test results experienced over the course of a screening regimen; (3) To use our newly developed statistical methods to analyze risk- based screening mammography regimens using 15 years of data from the Breast Cancer Surveillance Consortium (BCSC). Unbiased estimation of harms and benefits of repeat breast cancer screening tailored to women's individual breast cancer risk levels previously has not been possible because existing methods fail to account for important features of the observation scheme. Under Aim 3 we will analyze data from the BCSC using statistical methods developed under Aims 1 and 2 and will compare alternative breast cancer screening regimens tailored to women's breast cancer risk profiles. By contributing information about the potential harms and benefits of risk-based breast cancer screening regimens, this work will inform policy-makers and facilitate communication between patients and providers. This will aid both individual decision makers and development of evidence-based guidelines.