Statistical Methods for the Spatio-Temporal Assessment of Social Disparities in

社会差异时空评估的统计方法

• 批准号: 8589658
• 负责人: Brent Andrew Coull
• 金额: $ 14.78万
• 依托单位: HARVARD SCHOOL OF PUBLIC HEALTH
• 依托单位国家: 美国
• 财政年份: 2008
• 资助国家: 美国
• 起止时间: 2008-09-10 至 2018-06-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/8589658
• 关键词: Address; Advisory Committees; Algorithms; Anus; Area; California; Cancer Etiology; Censuses; Complex; Computer software; Data; Data Analyses; Databases; Development; Diagnosis; Ethnic Origin; Geographic Information Systems; Goals; Health Sciences; High Performance Computing; Human Resources; Incidence; Individual; Informatics; Instruction; Investigation; Joints; Knowledge; Lead; Link; Malignant Neoplasms; Mediation; Medical Surveillance; Methodology; Methods; Modeling; Monitor; National Cancer Institute; Population; Race; Research; Research Personnel; Resources; Role; SEER Program; Schedule; Small-Area Analysis; Socioeconomic Status; Source; Statistical Computing; Statistical Methods; Strategic Planning; Surveys; Time; Travel; Uncertainty; Validation; Work; anticancer research; burden of illness; cancer health disparity; cost; data management; health disparity; knowledge of results; malignant breast neoplasm; meetings; method development; mortality; multilevel analysis; neoplasm registry; novel; open source; programs; racial and ethnic; screening; social; social disparities; socioeconomics; spatiotemporal; surveillance data; therapy design; trend; user-friendly

This is a proposal to develop statistical methods for the effective analysis of social health disparities spanning the cancer continuum. Our proposed methods are motivated by and applied to investigation of US disparities in breast cancer incidence and survival, although the methods have wide applicability in other cancers. In spatio-temporal analyses of trends in cancer incidence, one source of uncertainty that can potentially have large impacts on analysis conclusions is small area intercensal population count uncertainty. Another important source of uncertainty in large cancer registries is the lack of individual-level data. This source of uncertainty yields conservative estimates ofthe annual costs of socioeconomic disparities, and better methods are needed to accurately estimate and ultimately reduce such disparities. Finally, popular methods for assessing the interplay of race/ethnicity and socioeconomic status on cancer disparities have the potential to yield severely biased inferences. To address these sources of uncertainty in the spatio- temporal analysis of cancer health disparities, we propose to develop new statistical methods and apply them to data from the Surveillance, Epidemiology, and End Results (SEER) database linked with data from the National Longitudinal Mortality Survey (NLMS). Specific aims ofthe project are to develop, evaluate, and implement (1) novel methods to incorporate intercensal population count uncertainty in the spatio-temporal analysis of cancer incidence data; (2) novel methods to integrate area- and individual-level data on socioeconomic position (SEP) to quantify socioeconomic disparities in cancer survival, accurately; (3) a causal inference framework to disentangle the roles of SEP and race in cancer survival disparities; and (4) fast algorithms and user-friendly open-source software for implementation ofthe developed methods. We will provide first-of-its-kind knowledge concerning the joint and individual impacts of race/ethnlcity, area and individual-level SEP on breast cancer incidence, mortality, and survival. Our newly developed methodology and resulting knowledge will contribute to key goals in NCI's and NIH's Strategic Plans to reduce and ultimately eliminate health disparities. RELEVANCE (See instructions): This project proposes to develop statistical methods for the spatio-temporal assessment of social disparities in cancer. The resulting methods will be applied to yield new evidence on the interplay between racial/ethnic and socioeconomic disparities in breast cancer incidence and survival. The project will make software available so that the developed methods are broadly applicable for routine analyses of other cancer surveillance data.

这是一项旨在发展统计方法以有效分析社会健康差距的建议 跨越了癌症的连续体。我们提出的方法的动机和应用调查美国 乳腺癌发病率和生存率的差异，虽然该方法在其他领域具有广泛的适用性， 癌的在对癌症发病率趋势的时空分析中，一个不确定性的来源， 可能对分析结论产生较大影响的是小区域两次普查间人口计数的不确定性。 在大型癌症登记中，另一个不确定性的重要来源是缺乏个人水平的数据。这 不确定性的来源产生了对社会经济差异的年度成本的保守估计， 需要更好的方法来准确估计并最终减少这种差异。最后，受欢迎 评估种族/民族和社会经济地位对癌症差异的相互作用的方法， 可能产生严重偏见的推论。为了解决空间中的这些不确定性来源， 癌症健康差异的时间分析，我们建议开发新的统计方法和应用 监测、流行病学和最终结果（SEER）数据库中的数据， 国家纵向死亡率调查（NLMS）该项目的具体目标是开发、评估和 实施（1）新方法，将两次人口普查之间的人口计数不确定性纳入时空 癌症发病率数据的分析;（2）整合区域和个人水平数据的新方法， 社会经济地位（SEP），以量化癌症生存的社会经济差异，准确;（3）a 因果推理框架，以解开SEP和种族在癌症生存差异中的作用;和（4） 快速算法和用户友好的开源软件，用于实现所开发的方法。 我们将提供有关种族/民族城市、地区的联合和个人影响的首创知识 以及个人水平的SEP对乳腺癌发病率、死亡率和生存率的影响。我们新开发 方法和由此产生的知识将有助于实现NCI和NIH战略计划的关键目标， 减少并最终消除健康差距。 相关性（参见说明）： 该项目提议制定社会差距时空评估的统计方法 在癌症中。由此产生的方法将被应用于产生新的证据之间的相互作用种族/族裔 和社会经济差异的影响。该项目将使软件 因此，所开发的方法广泛适用于其他癌症的常规分析。 监视数据。