April 10, 2009

Clinical records rich source for research

Featured Image

Dan Masys, M.D., talks about the Synthetic Derivative research database during a recent lecture at the School of Nursing. (photo by Susan Urmy)

Clinical records rich source for research

Researchers at Vanderbilt can now access data from clinical records representing more than 1.7 million unique individuals.

The records, which have had all identifying information removed, are part of the “Synthetic Derivative,” a database that has been built as a companion to BioVU — Vanderbilt's biobank of DNA samples linked to clinical records.

The Synthetic Derivative (SD) is a research resource in its own right, Dan Masys, M.D., said during a lecture this week at the School of Nursing.

“The SD allows a systematic query of events — such as 'how many patients do we see here who have this condition or who got that drug,'” said Masys, chair of Biomedical Informatics. “It is designed to support the detection of previously unsuspected co-morbidities and causal relationships.”

Many other academic medical centers have clinical research “warehouses,” Masys said, but those databases usually contain only structured values such as diagnostic codes, procedural codes, medication lists and laboratory values. In addition to structured values, the SD contains narratives including clinical notes, nursing notes, surgical reports and discharge summaries.

“This is where Vanderbilt is really standing out from other academic medical centers,” Masys said. “We include this very rich fabric of textual documents that represent the nuances of clinical care.”

The SD is available to Vanderbilt users who have a VUnetID and password and can be accessed from the Vanderbilt StarBRITE Web site (www.mc.vanderbilt.edu/starbrite/).

To use the SD, researchers must first obtain IRB approval — using a short form procedure — and sign a use agreement. A video tutorial and additional help information are available on the StarBRITE site.

The StarBRITE site also offers a new tool called the Record Counter that allows users to determine the number of patient records that would meet a study's demographic and diagnostic criteria. The Record Counter provides only the number of cases and limited demographic data, and users do not need prior approval.

For more information, contact Melissa Basford at 343-0121 or melissa.basford@vanderbilt.edu.