•  
  •  
 

Article Title

Leveraging the Virtual Data Warehouse to Support Implementation of Precision Medicine for Oncology at Catholic Health Initiatives

Publication Date

8-10-2017

Keywords

cancer, information technology, virtual data warehouse

Abstract

Background: Catholic Health Initiatives (CHI) is implementing a pilot program for precision medicine (PM) in oncology. The application to support this program integrates patients’ demographic, clinical, treatment and outcome data with molecular data from tumor testing into a central data store, which is used to inform personalized treatment options and match individuals with ongoing clinical trials. We assessed the extent to which the virtual data warehouse (VDW) could support the data needed for this initiative.

Methods: The CHI Institute for Research and Innovation data science team reviewed data requirements for the application with respect to content and format, and compared these to current VDW specifications. We identified data gaps between the application requirements and the VDW. We also identified data gaps within our own organization to support both of these and explored how we might leverage our partnership with the operations group through this initiative to facilitate acquisition of these data for mutual benefit.

Results: Of the 70 required data elements, 42 elements could be mapped to the VDW specifications (60%). We identified data gaps for patient observations such as functional performance scores, drug orders and administration, and narratives for procedures performed, which the VDW does not capture. The VDW also does not contain information on medications dispensed in inpatient settings. Significant gaps in data currently available at CHI to support the VDW and the PM application exist with pharmacy (medications) and infusion data. Results from this gap analysis have informed ongoing discussions with the PM operations team about the feasibility of acquiring these data, which would support both applications, and providing the funding through the PM initiative to do so. Further, we are discussing options for acquiring genomics data generated by PM into our data repository to be available for future research.

Conclusion: Our analysis suggests that the VDW can largely support the PM initiative, which would facilitate broad and efficient implementation of PM across multiple facilities. Data gaps that we identified inform potential need and opportunities for expanding the VDW to support PM initiatives. This effort highlights an example of a bidirectional partnership between research and health care operations, which is an essential component for learning health care systems.

Share

COinS