VUMC team creates COVID-19 research registryApr. 23, 2020, 10:38 AM
by Paul Govern
Out of the electronic health records (EHRs) of patients seen at Vanderbilt University Medical Center, a team in the Department of Biomedical Informatics is creating a COVID-19 patient registry as a platform for research.
“Over the next year there will be an urgent need to understand the biological underpinnings of COVID, involving researchers from almost every medical specialty,” said the leader for the registry project, Josh Peterson, MD, MPH, professor of Biomedical Informatics. “And one big-data approach is to aggregate cases from the EHR and curate key outcomes and phenotypes within the data, to make it easier for researchers to conduct more complex analyses.
“At VUMC, research efforts will include seeing if we can pair our COVID health record data with genetic data that we already have from BioVU, our DNA bio-bank,” Peterson said.
Through April 14, clinicians at VUMC assessment centers had evaluated 19,824 people for COVID-19 and had diagnosed the disease in 1,445 people.
The registry, which includes confirmed cases and people who have tested negative COVID-19, will continue to grow as more people are evaluated at VUMC.
According to Peterson, in the short term the registry potentially could help researchers shed light on questions such as which patients need to be hospitalized, which patients need a ventilator and which of those who test negative are at higher risk for later turning positive for COVID-19.
He and others are thinking about how best to pool the registry data (in anonymized form) with patient data from other research groups around the country, “to answer important questions with much larger numbers.”
As part of her work with the Center for Precision Medicine, Lisa Bastarache, MS, research assistant professor of Biomedical Informatics, is devising algorithms to automate selection of cases and controls for the registry. In an April 8 videoconference attended by more than 300 people, Bastarache described the early stages of this effort.
“I think that we’re in a situation where everybody is watching and is hoping for some answers. I think that people who do data science right now, and EHR-based data science, have enormous opportunities,” she said.
For registry inclusion criteria the team is currently relying on COVID-19 test results; as time goes on, other data will come into play in the selection algorithm, such as billing codes and clinical notes.
Peterson and Bastarache envision an all-purpose, longitudinal registry, replete with demographics, billing codes, lab results, diagnoses, signs and symptoms, vitals, co-morbidities, recent exposures (e.g., smoking), disease progression and long-term outcomes.
Through the eMERGE (Electronic Medical Records and Genomics) Network, where Peterson is principal investigator for the national coordinating center, plans are forming to pool COVID-19 EHR and genomics data from far and wide. The network includes VUMC, Harvard, Columbia, Mayo and 10 other institutions.
For COVID-19 guidance for the Vanderbilt research community, please visit the Vanderbilt Institute for Clinical and Translational Research.