A Multilevel Data Analytic Solution to Advance Population Cancer Research

Project Summary/Abstract The proposed work will develop a first-of-its-kind, transportable informatics tool, the Population Cancer Assessment and Surveillance Engine (Pop-CASE), that integrates patient-level cancer registry data for a population with granular community-level data in a standardized data model, while providing a user-friendly

query tool to facilitate quick searching by researchers and community outreach professionals. Broadening access to these types of data for researchers and other stakeholders will advance the pace and breadth of cancer disparities research and will catalyze action to reduce disparities. Building on the work of an existing prototype and with ongoing input from a nationwide Steering Committee, we

aim to 1) build the Pop-CASE data base and computational layer using an extended set of community and health system data sources and embedding additional calculation capabilities, 2) build a user-friendly interface with export capabilities, and 3) develop an Implementation kit featuring a software container for creating

location-specific instances of Pop-CASE and an implementation guide. These aims will be accomplished by first creating a relational data base in PostgreSQL with tables derived from North American Association of Central Cancer Registry (NAACCR) standard record formats; U.S. Census American Community Survey data; Health Professional Shortage Area (HPSA) data; small area estimates for

Behavioral Risk Factor Surveillance System (BRFSS) cancer screening, risk behavior, and comorbidity burden measures; National Provider Index (NPI) and Food and Drug Administration (FDA) provider and facility location information; historical “redlining” data reflecting discriminatory lending practices; and other sources as guided

by our Steering Committee. A computational layer with logic for age adjustment, distance calculations, and calculation of common indices of disparity or disadvantage will be coded in Python. A user interface built in a Javascript-based framework will enable specification of complex queries with output at various user-selected

geographic levels and options for report export—all with embedded capabilities for tiered access to small number counts and patient-level data sets. A modern container framework will be used to develop a “PopCASE-in-a-box” software container allowing other cancer centers or cancer registries with access to

patient-level registry data to create a secure, location-specific instance of Pop-CASE. In the final project year, the University of Southern California (USC) project team members will use the PopCASE container to implement a PopCASE instance based on Los Angeles County cancer registry data and known as “LA-CASE.”

PopCASE will provide a new tool for cancer control researchers, with place-contextualized cancer data at a fine geographic scale. It can also provide the Community Outreach and Engagement (COE) professionals and cancer center leaders working to curb cancer disparities at the local level with an unprecedentedly powerful

catchment area data tool.