To tell 23,840 people whether their genes, family history and clinical records add up to elevated risk for any of 11 common diseases, the eMERGE Network first had to build something to do the arithmetic.
A recent paper in JAMIA Open describes that system: the R4 (Recruitment, Results and Risk Reduction) Portal, built over 15 months by informatics teams at the Vanderbilt Institute for Clinical and Translational Research, or VICTR. The platform pulled together electronic health record data from 10 health systems across the nation, genetic results from two partner laboratories — the Broad Institute of MIT and Harvard, and Invitae, now part of Labcorp — family health history collected through a tool called MeTree developed at Duke University, and participants’ own survey answers. The platform then applied condition-specific logic to produce a clinical-grade, genome-informed risk assessment delivered to the participant, their clinician and their medical record, with risk graded as “high risk” or “not high risk.”
“Each of the 11 conditions had its own definition of high risk, using different combinations of monogenic results, polygenic scores, family history and clinical factors,” said first and corresponding author Jennifer Morse, MS, IT Project Manager at VICTR. “That meant roughly 600 fields of conditional logic to apply the right high-risk criteria for each condition, so patients and providers have the information they need to make informed decisions.”
The conditions included asthma, atrial fibrillation, breast cancer, chronic kidney disease, coronary heart disease, colorectal cancer, prostate cancer, hypercholesterolemia, type 1 diabetes and type 2 diabetes.
The team built R4 on REDCap, the research data capture platform developed and hosted at Vanderbilt Health and now used by more than 8,300 institutions worldwide. REDCap supplied electronic consent, multilingual surveys — participants completed them in English or Spanish — and site-level access controls.
Custom code and application programming interfaces handled the rest, including a link to a University of Cambridge tool that calculates lifetime breast cancer risk.
An initial two-month test at two sites was followed by beta testing at eight more, with site staff addressing issues along the way: missing data such as height, weight or age of onset; mismatches between what participants reported on surveys and what their family history record showed; and technical fixes to the logic itself. Those findings produced network-wide protocols for when to recontact a participant, when to pull a value from the medical record and when imputation was reasonable.
The eMERGE Network is funded by the National Human Genome Research Institute (NHGRI) and comprises Children’s Hospital of Philadelphia, Cincinnati Children’s Hospital Medical Center, Columbia University, Icahn School of Medicine at Mount Sinai, Mass General Brigham, Mayo Clinic, Northwestern University, University of Alabama at Birmingham, University of Washington, and Vanderbilt Health.
NHGRI has placed the network’s coordinating center at Vanderbilt Health since eMERGE began in 2007, and for the current phase Vanderbilt Health is joined in that role by the Broad Institute, with additional expertise from Duke University. Josh Peterson, MD, MPH, Professor and Chair of Biomedical Informatics, Vice President for Personalized Medicine, and one of two principal investigators for the eMERGE Coordinating Center, was the senior author on the recent study.
Other Vanderbilt Health authors on the paper are Megan He, PhD; Sophie Forman; Paul Harris, PhD; Adam Lewis, MS; Kyle McGuffin; Ellis Thomas; Travis Wilson; Georgia Wiesner, MD; and Jodell Linder, PhD, who contributed equally with Peterson as co-senior author.
Vanderbilt Health’s eMERGE Network site is led by Dan Roden, MD; Digna Velez Edwards, PhD, MS; and Wei-Qi Wei, PhD, MD.
The study was supported by National Institutes of Health grants U01HG011166 and U01HG011181, among others.
