Understanding disparities in cardiovascular toxicity among breast cancer survivors in Arkansas

Cardiovascular disease (CVD) is a rapidly growing public health concern for breast cancer (BC) survivors. This is due, in part, to cardiovascular (CV) toxicities from common cancer therapies that are associated with CVD. CV toxicities are more common in certain patient populations, including those from communities with economic challenges, those with limited insurance coverage, and those residing in rural areas or neighborhoods with limited resources. Yet, the complex role of demographic, economic, and geographic factors in influencing CV toxicity risk among BC survivors remains unknown. Few studies have examined how such factors interact to affect CV toxicity outcomes. Given the known relationship between demographic, economic, and geographic factors and cardiometabolic conditions, variation in these contexts may contribute to CV toxicities among BC survivors through comorbid cardiometabolic dysfunction. Therefore, to improve the CV health of BC survivors, it is critical to examine CV toxicity risk in the context of cardiometabolic dysfunction. To address this issue, the following specific aims will be completed: Identify the extent of differences in incident CV toxicity among BC survivors in Arkansas across key demographic, economic, and geographic factors. Develop a predictive algorithm for risk stratification in BC survivors at high risk for CV toxicity using machine learning approaches that incorporate demographic, economic, and geographic factors with traditional clinical characteristics. Data collected from 2013–2019 in the Arkansas All-Payer Claims Database (APCD) linked to the Cancer Registry, as well as clinical and refined demographic, economic, and geographic information from the electronic health records system at the University of Arkansas for Medical Sciences, will be utilized in this study. A longitudinal analysis for the development of CV toxicities in a cohort of BC survivors (stage I–III at diagnosis), with passive follow-up in the claims data through 2023, will be conducted. Machine-learning methods will be used to develop an algorithm that predicts CV toxicities among BC survivors based on demographic, economic, geographic, and other clinical factors. This K01 will: 1) provide training in contextual influences on health and health outcomes research; 2) promote research skills using large-scale, longitudinal administrative healthcare data; 3) develop competence in advanced analytic methods; and 4) increase understanding of BC survivorship and provide content expertise in cardio-oncology research. This study responds to the NHLBI’s compelling question (5.CQ.10) to reduce cardiac morbidity and mortality in cancer survivors. By identifying factors that contribute to differences in CVD outcomes among BC survivors and using them to predict CV toxicity, this research can inform targeted interventions (e.g., multidimensional intervention programs tailored to patients’ clinical profiles and community and environmental factors) to improve the CV health of this population. Modified