Data respository helps improve research speed, quality at Columbia

Tools

At Columbia University, development of a data repository and improved workflow have made outcomes research easier in the department of urology.

An article published at the Journal of American Medical Informatics Association describes the department's move from flat tables for data management to use of the open-source centralized research repository, along with a process involving a principal investigator, statistician and informatician.

Comparing pre- and post-test periods, the department's average annual retrospective study publication rate was 11.5 and 25.6, with the time to prepare retrospective studies falling from 12 months to less than six months. Publication impact scores grew from an average of 1.7 to 3.1, as well.

The department chose to use CAISIS, an open-source, web-based data management system that integrates research data with patient care data. It's the same technology on which Nova Scotia's breast health management system is based.

It required extensive work to eliminate redundancy in existing datasets and to map the data structure at Columbia, the researchers noted. Now a monthly feed of updated data keeps the system refreshed, including outpatient as well as inpatient information.

With the improvements in data quality and new workflow, research "has evolved from an ad hoc 'analyze now, ask questions later' process to a systematic one," the authors wrote. It also has allowed the department to expand its research staff, since multiple people now can access the datasets simultaneously.

Multiple studies have called into question the quality of data used for research, especially that gleaned from electronic records, prompting the American Health Information Management Association to call for better standards for EHR data governance.

Keith Marsolo of the biomedical informatics division of Cincinnati Children's Hospital Medical Center also detailed in a JAMIA article all the headaches his organization undertook and the workarounds required for EHR data to be used effectively in research. He joined the call for technology systems that make the sharing of data easier to better align research and clinical practice.

"If we are to ever achieve the vision of a learning health system, where learning occurs with every patient encounter, we need to ensure that the clinical information systems are configured to allow this learning," he wrote in the article.

To learn more:
- read the article

Related Articles:
Data discrepancies threaten research quality
AHIMA calls for better EHR data governance
Quality of EHR data too inconsistent for reuse in research
Open-source software suite aims to improve breast health management
EHR redesign required to better integrate research, practice