Skip to content Skip to navigation

Using big data to improve patient care

December 4, 2013
by Richard R. Rogoski
| Reprints

Initial results of a system-wide program to use data analytics for predictive modeling and to improve overall patient/resident care have been successful, Charlotte, N.C.-based Carolinas HealthCare System recently announced.

"In the end, healthcare analytics is about improving the health of the patient and of the population," said Michael Tarwater, the system's CEO, in a press release. "To better understand and address their needs, we have to be able to see the bigger picture, and as we continue growing our capabilities, we can help patients more quickly and efficiently."

The healthcare system, covering parts of both North Carolina and South Carolina, includes a wide variety of providers and facilities, including academic medical centers, hospitals, home health agencies, nursing homes, hospices, freestanding emergency departments and physician practices. It created a comprehensive enterprise data warehouse to store 10 terabytes of patient/resident-centered data related to its more than 10.5 million client encounters a year.

Using data-mining and analytics tools with its electronic health record system, the healthcare system has analyzed 40 different patient/resident variables deemed highly predictive of unplanned hospital readmissions and has achieved an 80 percent accuracy rate in predicting a patient's/resident's chances of being readmitted to the hospital within 30 days of discharge. On the outpatient side, Carolinas HealthCare System now is able to monitor patient/resident improvements system-wide, and by using what is known as geospotting, it can identify health needs and outcomes within specific segments of individual communities.