Design, extraction, and dashboard creation of an expandable recruitment data mart for diabetes clinical trials.

Bioinformatics Internship Presentation

Ariel Roane (Mentor: Dr. Ian M. Brooks, MedStar Health Research Institute)

August 29, 2017, 2:30pm, Room 341, Basic Science

Medstar Health Research Institute is the Research wing of Medstar Health, conducting translational and health science research on over 185 information systems. A major task of the various research groups within the organization is the selection of patients for clinical trials. Currently a solution through the CERNER EHR is still under development. In response, we developed an in-house solution consisting of a generic recruitment data mart and Tableau dashboards.

The database was designed with input from specialists within the diabetes research group as to the necessary variables for inclusion and exclusion of patients. These variables were then made generic and incorporated into the design of a dimensional data mart that could be expanded for other clinical trial recruitment strategies. The information was mapped and extracted/transformed/loaded from the CERNER EHR using custom SQL queries and python classes (based on connections through the cx_Oracle and pypyodbc modules). The Tableau dashboards were created for intuitive exclusion/inclusion of patient records, along with exportation of patient/provider contact information. The dashboards were elevated to the organizational Tableau server to allow researchers within the MedStar network to further validate the tool for its eventual use.