Measuring What Works

Health Data Management Magazine, December 1, 2009

Medicine is largely about trial and error. So you have to measure what works." That's how Audrey Fisher sums up why Sequoia Hospital is using data mining to conduct outcomes research in its cardiac care program.

For example, by using data analytics, the hospital detected a pattern that certain heart patients developed problems with bleeding during surgery if they received a particular blood clot prevention drug in the days before the procedure. The analysis led physicians to delay surgery for those who have been on the drug, says Fisher, the hospital's manager of cardiovascular outcomes and quality.

To help it accomplish these feats, the 345-bed community hospital in Redwood City, Calif., used specialized software from SPSS Inc., Chicago. Fisher and four part-time nurse coordinators on her team can run custom data queries for physicians that instantly yield results. And, she stresses, the queries are done without the benefit of a full-blown electronic health record.

Since 2000, the hospital has used the SPSS software to dive into data stored in a cardiac data repository from Heartbase Inc., Chicago. The repository automatically receives demographic information from the hospital's admission-discharge-transfer system. But the nurse coordinators manually review paper records and enter required information into the repository that, in turn, fuels powerful research.

Sequoia Hospital originally obtained the repository so it could gather data to submit for national quality improvement projects of the American College of Cardiology and the Society of Thoracic Surgeons. The goal was to receive benchmarking reports. However, by pairing the repository with the SPSS analytical software, the hospital "can slice and dice the data a million different ways," Fisher says.

The database, in place since 1994, now has information on 10,000 patients. It includes such details as how long the patient spent in the operating room, complications that developed, disease pathology that existed before the surgery, mortality rates, medications and much more.

"I can run queries whenever a surgeon wants to know something," Fisher explains. "I'll get calls on a Friday night from a physician asking for our mortality rates for a specific category of patients. This opens up a whole new world for us." One recent study determined that patients who have uncontrolled diabetes have much better outcomes if they spend a few weeks under the care of an endocrinologist getting treatment for the diabetes before heart surgery.

The hospital also can provide extremely specific insights to patients when discussing whether to proceed with their surgeries, including aortic valve replacements, one of Sequoia's specialties. "Because we have so many thousands of patients in the database, when a patient comes in to see a physician, we can take into account their risk factors and we can predict the risk of a stroke if we do the surgery," Fisher says.

What makes the analysis of data so powerful, Fisher says, is that "it's not just the national rate for complications; it's our experience with local patients."

A recent report from the Society for Thoracic Surgeons determined that Sequoia Hospital was the highest scoring hospital in the nation from 2004 through 2008 for its aortic valve replacement mortality rate. Fisher attributes this performance, in part, to the use of data to help doctors create protocols that yield the best results.

The hospital's mortality rate for cardiac surgery, without adjustments for the severity of the cases involved, declined to 1.7% in 2008 from 3.8% in 2003.

(c) 2009 Health Data Management and SourceMedia, Inc. All Rights Reserved.

http://www.healthdatamanagment.com/ http://www.sourcemedia.com/

For more information on related topics, visit the following channels:

• Business Intelligence
• Decision Support
• Electronic Health Records
• Hospitals