Multilevel Mediation Analysis to Explore Racial Disparities in Breast Cancer Recurrence and Survival using CDC Special Studies

乳腺癌是最常见的癌症，也是癌症死亡的第二大原因 在所有种族的美国女性中。尽管复发率和生存率有所改善， 在美国，白色和黑人女性之间的乳腺癌发病率仍然存在显著差异。 以前的研究发现，更先进的和侵略性的肿瘤和低于最佳 治疗可能解释了黑人女性复发率较高和生存率较低的原因， 与白色女性相比。然而，据我们所知，没有任何研究 共同考虑潜在因素，不仅包括个人层面的信息， 治疗历史和肿瘤特征，而且还提供者特征，以及身体和 社会环境信息。由于目前分析方法的局限性和缺乏 综合数据集，研究人员还无法区分每一个相对影响 一个因素造成了不同层次（人口普查区或个人）的总体种族差异 水平）。 我们将利用疾病预防控制中心资助的特别研究，包括护理模式研究， 比较有效性研究（CER）和以患者为中心的癌症登记数据 结果研究（PCOR），以检查乳腺癌种族差异的决定因素 生存和复发使用一种新的多层次中介分析。CER和PCOR 研究收集了女性乳腺癌的完整护理标准和肿瘤特征信息 癌症患者。路易斯安那州肿瘤登记处参与了所有这些研究，并收集了 路易斯安那州乳腺癌患者的信息。我们计划将数据集与 人口普查数据和许多其他数据来源，以获得环境和供应商信息。 我们提出了一种新的统计方法，可以共同考虑潜在的相关 多个介质，并区分间接影响，从不同的个人因素 可能有助于种族-乳腺癌生存和复发关系的水平。如果 成功的，这一建议将提供最全面的探索种族差异， 美国乳腺癌患者的复发率和生存率。

Breast cancer is the most common cancer and the second leading cause of cancer death among American women of all races. Despite improvement of recurrence and survival rates of breast cancer in the United States, significant differences between white and black women remain. Previous studies have found that more advanced and aggressive tumors and less than optimal treatment may explain the higher recurrence and lower survival rates for black women as compared to white women. However, to the best of our knowledge, there has been no research that jointly considers potential factors that include not only individual level information such as treatment history and tumor characteristics, but also provider characteristics, and physical and social environmental information. Due to limitations of current analytic methods and the lack of comprehensive data sets, researchers have not been able to differentiate the relative effect each factor contributes to the overall racial disparities at different levels (census tract or individual levels). We will use the CDC-funded special studies, including Patterns of Care study, Enhancing Cancer Registry Data for Comparative Effectiveness Research (CER), and the Patient Centered Outcomes Research (PCOR), to examine the determinants of racial disparities in breast cancer survival and recurrence using a novel multilevel mediation analysis. The PoC, CER, and PCOR studies collected complete standard of care and tumor characteristic information for female breast cancer patients. The Louisiana Tumor Registry participated in all these studies and collected the required information from Louisiana breast cancer patients. We plan to link the datasets with census data and many other data sources to obtain environmental and provider information. We propose a novel statistical method that can jointly consider potentially correlated multiple mediators, and differentiate the indirect effect from every individual factor from different levels that may contribute to the race-breast cancer survival and recurrence relationship. If successful, this proposal will provide the most comprehensive exploration of racial disparities in breast cancer patients’ recurrence and survival in the United States.