KM in healthcare: Distributed networks boost clinical research

Conducting large-scale clinical research is getting easier and faster for healthcare organizations due to the development of distributed research networks. The newly established networks are helping to overcome a major knowledge management hurdle in healthcare by allowing researchers to send queries to the electronic data sets of collaborating health systems, which run the query against their data and send back results.

In a March 2018 presentation about the National Institutes of Health Collaboratory Distributed Research Network (DRN), Prof. Lesley Curtis, director of the Center for Population Health Sciences at the Duke School of Medicine, described how the distributed research network model has been applied to medical product safety surveillance in a U.S. Food and Drug Administration (FDA) effort called Sentinel, in which queries go from the organization making the query to the distributed network, and then curated results go back to the requestor. The partners agree to use the same data model, which eliminates the need to pull the data into one data warehouse or to share confidential or proprietary data, she said.

Here is a brief deep dive into one example that takes advantage of the Sentinel network: a study called “Impact AFib.” The aim is to increase the use of oral anticoagulation (OAC) among patients with AFib and at risk of stroke. In a presentation last year to the NIH Collaboratory, Sean Pokorney, M.D., M.B.A., a cardiologist and member of the Duke Clinical Research Institute, said that preliminary research of claims data suggested that among the population with AFib and stroke risk factors, only 52 percent were being treated with anticoagulants.

“The underuse of anticoagulation is a major public health priority and a priority for health plans as well. They want to deliver optimal care,” Pokorney said.

Oral anticoagulation study

The study designed educational mailings about OAC for patients and their providers, with the help of patient representatives. The trial uses the FDA’s Sentinel database and claims data from five large commercial insurers to identify cohorts of patients.

“We can write queries against the data and what gets returned is aggregated data from each site. It translates into an incredibly robust and efficient system with rapid analytic capability,” said Noelle Cocoros, DSc, MPH, an epidemiologist and research associate in the Department of Population Medicine at Harvard Medical School and Harvard Pilgrim Health Care Institute.

The IMPACT team randomized 80,000 patients with AFib diagnoses and stroke risk into an early intervention group and a delayed intervention group, who will get the same notices a year later. A commercial institutional review board (IRB) waived the need to consent patients into the study. The trial will also assess outcomes associated with the treatment, including stroke.

“That same kind of querying and results production can inform what we know about the quality of care, and it can inform public health surveillance,” Curtis said. “Certainly, it can support randomized trials, comparative effectiveness research and a wide variety of clinical research.”