Study to search EHR for undiagnosed genetic diseasesOct. 13, 2022, 9:30 AM
by Paul Govern
With a five-year, $4.3 million grant from the National Human Genome Research Institute, part of the National Institutes of Health, a team at Vanderbilt University Medical Center will explore computational solutions to help address the problem of missed or delayed diagnosis for patients with rare genetic diseases.
Previous work by the project’s principal investigators, data scientists Lisa Bastarache, MS, and Douglas Ruderfer, PhD, has demonstrated the capability of computer algorithms to detect patients with genetic disease based on characteristic patterns of symptoms mined from the electronic health record (EHR). This suggests that programs may be developed both to help find patients who remain undiagnosed and to identify factors that contribute to diagnostic delay.
“Diagnosing genetic diseases can be a huge challenge for clinicians,” said Bastarache, research associate professor of Biomedical Informatics.
“Some patients wait years or even decades before they are diagnosed. This is an area where data scientists might be able to assist clinicians by taking a systemwide approach to the problem.”
The researchers will start with a previously curated database of over 10,000 patients with confirmed genetic diseases. They plan to combine these real-world EHR data with information from the Online Mendelian Inheritance in Man, a source for authoritative clinical descriptions of genetic diseases. Using whole exome sequencing to test predictions, the team will also seek to gain an understanding of what leads to diagnostic delay.
“By understanding the clinical course of patients with rare genetic diseases through the health care system, and integrating with existing expert curation, we will have a more complete picture to assist providers in earlier recognition and diagnosis of these patients,” said Ruderfer, associate professor of Medicine.
Joining Bastarache and Ruderfer for the project are Josh Peterson, MD, MPH, Julie Bastarache, MD, Kathryn Dahir, MD, and Rizwan Hamid, MD, PhD. The project is supported by NIH grant HG012657.