KM in healthcare: Distributed networks boost clinical research

ADAPTABLE’s goal is to enroll 20,000 patients. For that project, the Mid-South Clinical Data Research Network (CDRN), which has data on 9 million patients, takes electronic health records (EHR) data from health systems and transforms it into a common data model to run queries against. Researchers identified patients with coronary artery disease through phenotype and consented them electronically. Patients are being randomized into low-dose or regular aspirin groups and followed electronically for two years by pulling their EHR data and surveying patients as well. The goal is to see what the optimal dose of aspirin is.

“We have been recommending aspirin for decades and still don’t know what the optimal dose is,” Russell Rothman, M.D., VP for population health research at Vanderbilt University Medical Center, told an NIH Collaboratory meeting last year. “We are excited to be answering that question in a pragmatic way.”

In 2017, the NIH Collaboratory DRN solicited research queries to run against its data. It selected five research projects:

♦ Incidence and recurrence of hepatocellular carcinoma associated with oral direct-acting antivirals,

♦ Identifying chemotherapy-induced peripheral neuropathy (CIPN) and its treatment,

♦ Antibiotic dispensing in emergency departments and ambulatory settings,

♦ Estimating opioid users and diagnoses of opioid use disorder and opioid overdose, and

♦ Estimating prevalent long-term bisphosphonate use.

Curtis stressed that there are many advantages to the DRN approach, including the ability to work with analysis-ready datasets that cover millions of individuals. “These data are standardized according to a common data model. The queries and responses rely on an audited and secure platform called PopMedNet,” she said. “There is a large and growing suite of analytic tools that support these analyses, ranging from simple descriptive analyses to much more complex comparative analyses. In the end, the key advantage is this approach supports and really enables efficient multisite studies.”