Mechanisms Underlying SES Disparities in Breast Cancer Mortality

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Socioeconomic (SES) disparities in survival from breast cancer developed during the 1980's and 1990's, and have persisted into the early 2000's. The existing literature and/or our preliminary data provide evidence of process disparities among SES groups in the use of screening and follow-up mammography, the use of breast conserving surgery with radiotherapy, the choice and use of alternative adjuvant endocrine therapy (aromatase inhibitors vs. tamoxifen) and possibly persistence with endocrine therapy once initiated. There is also evidence that SES disparities may exist among breast cancer patients in the use of low-volume hospitals, which are associated with worse survival. What is unclear is the extent to which these process factors inter-relate to lead to the observed survival disparities, and which public policy or health care delivery interventions deserve greatest attention in a program designed to reduce these disparities. In this application, we propose to show empirically the extent to which each of these factors contribute to the observed SES disparities in mortality among women with breast cancer. Specifically, our aims are to: (1) To document SES disparities in patterns of breast cancer care, and in 3- and 5-year mortality using a contemporary cohort of elderly women with incident breast cancer; (2) To decompose the factors underlying SES disparities in breast cancer mortality, most notably, (i) screening, (ii) extent of disease; (iii) initial treatment; (iv) adjuvant and follow-up care; and (v) health system structural and policy factors; and (3) To simulate the effect of alternative public policies on cancer mortality and their implications for SES disparities therein. These aims will be carried out by analyzing data from the 2005-2007 SEER-Medicare-Part D data and followed until 2012. We propose to first identify patterns of breast cancer care, defined over the 60-month post-surgery treatment continuum, using the Classification and Regression Tree (CART) methodology. We will then apportion the observed SES disparities in mortality according to the contribution of each structural and process measures comprising these patterns of care, using the Oaxaca-Blinder regression-based decomposition approach, quantifying the portion explained by each, and the differences that remain unexplained. Finally, we will use parameter estimates generated by these analyses to simulate the anticipated effect of alternative proposals currently under consideration (e.g., increasing Part D premiums and co-pays, eliminating the coverage gap) in offsetting or exacerbating disparities in breast cancer treatments and outcomes. The proposed study will not only provide mechanistic information to focus efforts to remediate SES disparities in breast cancer mortality, but will also provide an innovative policy simulation that will inform efforts to determine how best to allocate the resources of public medical and prescription drug insurance programs in order to obtain the greatest value.