Crusade for Cancer Data: How a Non-SEER Site Populated the Virtual Data Warehouse Tumor Table
cancer, virtual data warehouse
Background: The virtual data warehouse (VDW) tumor table is a valuable data resource available to Cancer Research Network (CRN) researchers to examine the feasibility of potential projects or to conduct research. The Surveillance, Epidemiology and End-Results (SEER) registry provides timely and efficient data for VDW tumor tables. HealthPartners Institute (HPI) is a non-SEER site that has developed alternative approaches (claims-based algorithms, medical record review) to obtain tumor data and participate in cancer-related research. These approaches can identify false-positive cases, be time-consuming and be costly.
Methods: Over the past 15 years, HPI has undertaken numerous activities to identify a viable electronic data source to populate the VDW tumor table and thereby more readily participate in multisite studies within the CRN. These activities included two internally-funded capacity-building projects. The first, conducted in 2009, examined the ability to connect with the population-based cancer registry Minnesota Cancer Surveillance System, maintained by the Minnesota Department of Health (MDH), to identify cancer cases with specific diagnostic and treatment criteria preparatory to research. In 2011, our second project explored linking with the electronic registry system (ERS) at an HPI-owned hospital, Regions. More recently, HPI increased its tumor registry case ascertainment with the inclusion of care centers in Eastern Minnesota and Western Wisconsin as well as the merging of another health system, Park Nicollet.
Results: The 2009 feasibility project demonstrated that MDH was able to link 79% of cases identified by HPI claims data. Nonmatches occurred from misclassification by HPI-created algorithms, patients not living in Minnesota, or patients with cancers different from those identified by HP records. Concordance for determining eligibility (stage, date of diagnosis) was high, but not 100%. The 2011 feasibility project demonstrated that HPI programmers can access Regions cancer registry data directly via the ERS similar to other data sources (eg, Clarity, Epic) through procurement of a software license and training. The project programmer extracted registry cases from ERS, mapped data elements to variables outlined in the VDW tumor tables and performed quality assurance checks provided by the VDW tumor work group. Thus, ERS data is currently our electronic source for the VDW tumor table. Following the expansion of the HPI care network to include Lakeview Hospital and three Cancer Center of Western Wisconsin sites (Westfields Hospital and Clinic, Amery Hospital and Clinic, Hudson Hospital and Clinic) along with the 2014 merger of HPI and Park Nicollet organizations, HPI is currently extracting tumor data from the ERS at these sites and Park Nicollet Methodist Hospital. Ongoing efforts to enhance the HPI tumor file include obtaining any available tumor information from MDH for all HPI patients and members.
Conclusion: After time and vigorous exploration, HPI’s tumor data is part of the CRN Cancer Counter. More recent activities have enriched this data. HPI is in a better position to not only conduct internal cancer research, but also to participate in multisite studies. A visual timeline of the activities undertaken to identify and connect with tumor data sources, as well as the challenges, successes and proposed future work, will be presented.
Pawloski PA, Jackson J, Butani, DeFor T, Kane S. Crusade for cancer data: how a non-SEER site populated the virtual data warehouse tumor table. J Patient Cent Res Rev. 2017;4:175-6.