In this supplement application to U01 CA164975 under PAR-18-748 ?Research Supplements to Promote Data Sharing in Cancer Epidemiology Studies?, we propose to prepare, document, and submit cancer incidence, mortality, and case-fatality data from the Atherosclerosis Risk in Communities study (ARIC) to NHLBI's BioLINCC, a controlled-access database that contains the data from the parent ARIC study. Under the Cohort Infrastructure mechanism, we enhanced ARIC's infrastructure making it a bona fide Cancer Epidemiology Cohort with 4,743 incident primary cancer cases and 1,660 cancer deaths in 1987-2012. ARIC is a unique cohort of 15,792 White and Black adults that began in 1987 when participants 45-64 years old were recruited to be sex (55% female), race (27% Black), and geography (Forsyth Co., NC; Jackson, MS; Minneapolis, MN; Washington Co., MD) diverse. ARIC's strengths for cancer research include measured anthropometric, lifestyle, and medical data in clinical visits; calls to update data; repeated biospecimens and biomarkers; a GWAS all consenting participants and other `omics on large subsets; linkage to Medicare (CMS) claims data; and 30-year follow-up. Relevant to cancer survivorship, ARIC has extensively assessed cognitive and physical function. As products of the U01, we generated an analytic case file for total cancer incidence and mortality, and site-specific analytic case files for bladder, breast, colorectal, lung, pancreas, and prostate cancer that contain site-specific variables to characterize the tumor, follow-up time variables, and covariates needed to conduct high-quality cancer epidemiology analyses. For breast, colorectal, lung, and prostate cancers, the analytic case files also include adjudicated case fatality. We prepared detailed documentation for each file and developed sample SAS code and a macro that allows users to conduct visit-specific exposure updated survival analyses. We documented the utility of these cancer data for ARIC Cancer's broad research agenda, for collaborators, and consortia, by publications, conference abstracts, and approved ancillary studies and manuscript proposals. To maximize ARIC Cancer's value to the scientific community independent of ARIC and ARIC Cancer, while preserving confidentiality and being compliant with federal and state laws and regulations, we propose to modify the datasets and documentation we prepared as follows: 1) remove data that cannot be shared at an individual level, 2) remove or modify individual data that when crossed with ARIC data already submitted to BioLINCC potentially could lead to participant identification, 3) de-identify the data, 4) revise existing documentation to align with the revised data, 5) test the veracity of the de-identified data and documentation by performing the statistical analysis of a previously approved study and comparing its results to those when using the original dataset and documentation, 6) make final revisions to produce the final dataset and documentation, and 7) provide these materials to the ARIC Coordinating Center for uploading into BioLINCC as they do with the parent ARIC data. The work will be completed in 6 months.
By making the cancer incidence, mortality, and case-fatality data from the Atherosclerosis Risk in Communities study available in the same controlled-access database as the parent ARIC data, the reach of these cancer data is extended to independent investigators, including cancer epidemiologists and scientists from other disciplines and chronic diseases, who may have innovative ideas about the causes of cancer and poor cancer outcomes, or about how cancer may influence other outcomes.
Showing the most recent 10 out of 13 publications
