Skip to Content

Rapid, Self-serve Clinical Analytics: A Web and Hadoop-based Clinical Research Database (CRDB)

Project Description: The purpose of this project was to create a large-scale, easily accessed clinical research database (CRDB) with de-identified data from Loyola University Health System’s Epic electronic medical record (EMR). By significantly reducing the time and effort required for data identification and acquisition the self-service CRDB was developed to facilitate the institution’s clinical research and quality improvement initiatives, as well as educational programs for medical students, residents, fellows and other health professions trainees involved in research. To support this project, information technologies (IT) and experienced clinical research teams worked in concert to: 1. Develop a suitable technical platform for the clinical research database. “Big Data” Hadoop technology was selected as the database and analytic platform; 2. Define a series of simplified and refined clinical data structures; 3. Create and validate the required logic and routines to load the new repository with de-identified clinical information from Epic; 4. Create a supporting web site with predefined dashboards and ad-hoc query tools that allow end-users to directly identify targeted patient cohorts; and 5. Define and refine data categories, and identify frequent types of data requests.
Loyola University Chicago Stritch School of Medicine (SSOM) and the affiliated Loyola University Health System (LUHS) are located 15 miles west of downtown Chicago. LUHS has a mature Epic environment – implementation began in 2003. While the institution has more than 90 percent of clinical operations in this environment, it is not suitable for easy-access, large-scale clinical research studies. Prior to development of this CRDB, data identification and extraction from Epic was a complex and labor-intensive process, as well as an inefficient use of staff time and resources. Even mundane ‘Preparatory for Research’ analyses required significant staff effort, limiting throughput. Time from initial request to extraction often took more than two weeks with complex extractions taking one to two months. Post CRDB implementation, most routine requests are handled online via a self-serve web site.

Primary contact: Richard H. Kennedy, Ph.D., Vice Provost and Senior Associate Dean for Research
Institution: Loyola University Chicago Stritch School of Medicine

Join the ROCC Community

Receive updates on the Research on Care Community.

First Name:
Last Name:
Job Title:

Learn about AAMC medical research initiatives and publications.