Skip left side navigation and go to content
Epidemiologist Immerses Himself in Big Data as He Studies the Link Between HIV and Cardiovascular Disease
Gary H. Gibbons - December 9, 2013
As web-surfing consumers, we know that potentially every online search or purchase is being tracked, mined and analyzed as ‘big data’ that retailers use to optimize and customize their marketing pitch to us. As physicians, we gather information about our patients all the time, at every visit. And like major online retailers, we should be learning something about our patients from all of the information we gather, gleaning insights into how to better treat not only that particular patient, but all patients. However, many would argue that not only are we doing an inadequate job at capturing that information, but also at mining the information we do gather to improve care.
Researcher Matthew Freiberg, MD, MSc, an associate professor of medicine and epidemiology at University of Pittsburgh and a member of NHLBI’s HIV-CVD Collaborative, is working to change that.
As part of his research with the HIV-CVD Collaborative – and through other NHLBI-funded grants and awards – Matt is part of a team that has created a tool that will sift through the Veteran’s Administration’s electronic medical records (EMRs) looking for components of cardiac structure and function. The tool, which has been validated against manual chart review as being 94-95% accurate, will allow researchers to identify heart failure cases much faster and more efficiently than ever before. And while that tool was created to look for one particular piece of information specifically within the VA’s EMRs, it’s not a big leap to move from this one project, to expanding the tool’s applicability to other EMRs and other diseases, including rare diseases that are often difficult to study because of the challenges involved with finding enough patients who suffer from a particular condition.
Dr. Freiberg discusses the role of Big Data in his work and the scientific potential of Big Data. (12 MB, 15:05) Transcript
HIV and Cardiovascular Disease
Matt didn’t just help create the tool. He’s also analyzing the data that the tool extracts from the VA’s medical records. Among his more intriguing findings is the belief that HIV may carry an excess risk of 50% for having heart attacks. And even for those who manage to keep their viral load at less than 500 consistently for multiple years, there is still an excess risk of almost 39% for having heart attacks.
Dr. Freiberg discusses some of the most intriguing findings from his Big Data work. (3 MB, 4:02) Transcript
Dr. Freiberg discusses inflammation in relation to HIV and cardiovascular disease. (5 MB, 7:22) Transcript
Much of Matt’s thought-provoking work hinges on the data he mines using the tool he and his team created. But despite the success so far with this particular tool, many challenges still lie ahead of us when it comes to the intersection of science, healthcare and Big Data, not the least of which is the challenge of EMRs that don’t talk to each other.
However, he’s also optimistic – as am I and others at NIH – that the scientific opportunities inherent in the use of Big Data will help us all to work together to overcome the challenges and make the best use of all the data that is out there.
Dr. Freiberg talks about the challenges and opportunities that lie ahead for Big Data. (8 MB, 10:33) Transcript
But even more exciting for the individual researcher, the information he’s managing to extract from the VA’s EMRs has the potential to direct not just his line of research, but also that of other investigators who may find that the data he’s sharing lead them in new directions.
Much of Dr. Freiberg's funding comes via the NHLBI AIDS Program, which provides global leadership for research, training, and education programs to promote research in the areas of HIV-related heart, lung, and blood diseases, as well as blood-based therapies. The NHLBI AIDS Program recently issued new funding notices and notices of intent for upcoming funding opportunities. Learn more.