Abstract A006: Multi-modal estimation of causal influences of environmental agents on colorectal cancer in an understudied population

Abstract

Abstract Background: While there is a well-established causal relationship between several individual factors and contextual socioeconomic status with colorectal cancer, there is still a lack of strong evidence that the prevalence of environmental agents is independently associated with colorectal cancer in various geographical areas. Population-based studies assessing the causal relationship between environmental agents and colorectal cancer are limited by the scarcity of comprehensive data. Most previous studies have relied on the recall of self-reported exposures. Objective: In this study, we provide new and important insights by integrating publicly available ‘big data’ resources with hospital data to identify causal influences on colorectal cancer in an understudied rural population in Virginia. Methods: Data linkage was achieved by geocoding patients zip codes at diagnosis and spatially assigning contextual and environmental data to the hospital cancer records. Machine learning imputation methods were used to fill in the missing values that resulted from the linkage. A “Big Data” science approach was used to develop a multi-scale modeling of individual and contextual data with national and local exposomes from environmental exposures sources. The methodology is based on training a Bayesian causal network to better understand the causal inference from individual and contextual level data, as well as environmental exposures data on the outcomes of colorectal cancer. The primary source of data for this project is the Sentara Cancer Registry (SCR). The SCR collects data on cancer cases from eight Sentara hospitals. The information collected by SCR includes patient characteristics and clinical outcomes. The secondary source of data come from the U.S. Census Bureau (Census 2010). This data source is used to determine zip-code level socioeconomic status of each colorectal cancer case. Environmental data such as water quality, Superfund site locations, toxic release sites, and heavy metals were obtained from the Virginia Department of Environmental Quality. Environmental data was supplemented with Toxics Release Inventory (TRI) database, considered to be the most comprehensive data source on industrial toxic emissions in the US. Conclusion: This study contributed to the scientific innovation of leveraging mixed-methods data collection by building a comprehensive database for cancer research in an understudied population. Citation Format: Hadiza Galadima, Georges Adunlin, James Blando. Multi-modal estimation of causal influences of environmental agents on colorectal cancer in an understudied population [abstract]. In: Proceedings of the Twelfth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2019 Sep 20-23; San Francisco, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(6 Suppl_2):Abstract nr A006.