Using “big data” to improve patient care: Researchers explore a-fib treatments

TRACIE WHITE
August 10, 2015

A Stanford cardiac electrophysiologist and colleagues have used a unique research method to learn more about atrial fibrillation. Mintu Turakhia, MD, and collaborators at Medtronic and Massachusetts General Hospital, extracted data out of decades of continuously recorded medical information from implanted medical devices—pacemakers and defibrillators—in 10,000 heart patients. Then they linked it to medical records, and analyzed it.

The researchers’ goal was to explore whether patients who experienced sudden attacks of a-fib, an irregular and rapid heart rate caused by spasms of the heart’s upper chambers, should be treated with long-term anticoagulants like those who had permanent a-fib or whether perhaps temporary drug therapy could be considered an option. They wanted to know if a patient’s risk of stroke changes as a-fib comes and goes.

The results, which were published recently in Circulation: Arrhythmia and Electrophysiology, found that patients were at an increased risk of stroke the first seven days after their hearts went into a-fib.

A-fib, which afflicts more than 3 million Americans, is known to increase a patient’s risk of stroke – but exactly when this risk occurs is controversial. Currently, physicians recommend long-term anticoagulation for patients, whether the a-fib occurs in sudden attacks or is continuous. This study indicates that transient use of anticoagulants could be an option for some patients and deserves further investigation. Future treatment plans might explore the idea of some kind of wearable device that shows when a patient goes in and out of a-fib, then taking medications just when needed rather than for a lifetime, said Turakhia.

Turakhia told me the study also provides an important example of how using “big data” research methods can ultimately lead to improved clinical care. In an email, he explained:

This is truly a big data approach where we took raw data from implanted pacemakers and implanted defibrillators and linked it to clinical data. The medical device data comes from home remote monitoring systems that patients have and goes to the cloud. We pulled the raw data off the cloud and linked it to VA (Veterans Affairs) electronic health records, VA claims, Medicare claims, and death records. This is truly a novel approach where we are assembling highly disparate data sources and linking them to gain insight into disease.

Original article appeared here.

Learn more about the Campaign for Stanford Medicine's revolutionary Biomedical Data Science Initiative.