Front-runners say clinical data can help clinicians meet core measures
Even before the proposed rule on meaningful use of electronic health records was issued, leading-edge hospitals were looking for ways to mine clinical data and foster better patient outcomes. That now is even more of a priority as the government's plan to invest in information technology becomes more clear.
Hospitals have always had access to surrogate clinical data through billing systems. Many clinical decision support systems rely on billing-level data to drive physician-practice profiles and length-of-stay analysis.
"Now, with better tools emerging, many hospitals are beginning to mine transaction-level data to diagnose, plan and manage care," says Christopher McManus, lead health care IT partner at Tatum Healthcare Consulting, Atlanta.
"Data mining will become the most important function in the hospital," predicts William Fera, M.D., vice president of medical technologies, University of Pittsburgh Medical Center.
Fera says hospitals will have to create more reports for such things as the Centers for Medicare & Medicaid Services' core measures, making automated reporting mandatory. Real-time, predictive analytics will evolve in place of retrospective reporting.
"If I run a report to see what happened in my hospital today, it may be too late to change outcomes," Fera says. "However, if I'm alerted as soon as a patient is at risk or is not satisfying core measures, I'm in a position to affect outcomes."
Some emerging data-mining systems now allow users to run spontaneous reports.
Sequoia Hospital's cardiac surgery department mines data to feed a predictive analytics system from IBM-SPSS to determine if patients are appropriate surgery candidates, and to guide case management by creating protocols that yield best outcomes.
"In seconds, cardiac surgeons can assess the risk of mortality for a coronary bypass on an 80-year-old patient with recent history of heart attack," says Audrey Fisher, cardiovascular outcomes and quality manager. She says SPSS has helped the Redwood City, Calif., facility cut its cardiac mortality rate in half in five years.
The biggest challenge for most hospitals is gathering enough data.
"Hospitals with CPOE, and nursing and physician documentation in place, are well-positioned to generate significant data-mining benefits," says Mike Davis, executive vice president, HIMSS Analytics. Davis says more hospitals are using data mining to identify and avoid potentially negative outcomes, such as using protocols to monitor patients at risk for methicillin-resistant Staphylococcus aureus (MRSA) and catheter-associated urinary tract infections.
But data mining isn't without risk. Using medical data for marketing purposes is one hazard. The Federal Trade Commission has landed hard on entities that sell customer information without appropriate permissions. The Data Accountability and Trust Act of 2009 ups the ante on penalties.
"Hospital executives should be extremely cautious if approached to share EMR information for marketing purposes," says Elise Dieterich, partner, Sullivan & Worcester.
This article first appeared in the February 2010 issue of H&HN magazine.