Dr. Sweeney Culls Best Practices to Reduce Hospital Readmissions

February 2012

The following feature appeared in the Winter 2012 edition of Emory Health magazine, and was written by Sylvia Wrobel

John Sweeney, Emory's chief of gastrointestinal surgery, believes that each patient requires a physician's full, individualized attention—what he calls "looking the patient in the eye."

Patients love his approach. But when surgery’s clinical quality and patient safety program wanted to understand why some surgery patients were readmitted within 30 days of discharge, Sweeney chose a research partner who never lays eyes on a patient: James Cox, Georgia Research Alliance Eminent Scholar and director of the Experimental Economics Center of the Andrew Young School of Policy Studies at Georgia State.

Unplanned hospital readmissions are a big problem in health care, and not just for the patient. Although percentages differ, depending on conditions and hospitals, roughly six of every 100 patients discharged from hospitals nationwide find themselves back in the hospital within 30 days. Readmission for Medicare patients alone costs more than $17 billion annually. Keeping patients in the hospital longer lowers risk of complications leading to readmission, but unnecessarily prolonged stays lower quality of care and precipitate higher costs.

Sweeney and his surgeons initially approached the problem by focusing attention on individual patients (what went wrong?) and on which operations most often required readmission (those for pancreas, colon, and liver). Then, encouraged by Chris Larsen—Whitehead Professor and Chair of Surgery, who himself is deeply involved in health services research (HSR) to understand and improve outcomes for post-transplantation patients—Sweeney met Cox, who went straight for the numbers. He and his team of experimental economists analyzed hundreds of thousands of observations on more than 3,000 Emory surgery patients. The team of economists delved through vital signs, laboratory values, number and type of X-rays ordered, and other data recorded several times a day throughout each hospital stay. They tracked underlying medical problems. They matched zip codes to census data to come up with median income and the number of people per household in the census tract in which the patient lived.

"It was like the baseball movie, Moneyball," says Sweeney. "The economists threw out our preconceived notions and analyzed mountains of our data in whole new ways. They discovered things in mathematical trends that we couldn't see in any one patient."

For example, one of the traditional decisions for discharge depends on a normal white blood count the morning of release from the hospital. But the data showed that measuring how long the white blood count had been normal was important. The same with how long the patient had been following a normal diet.

And like all HSR research, the analysis went past clinical factors to look at behavioral, cultural, and social ones, such as whether patients had strong social support. Those who lived in census tracts with more people in each house were less likely to be readmitted.

As a profile emerged for risk associated with complications and unplanned readmissions, Sweeney and Cox developed a software tool that recommends whether an individual should be discharged or stay in hospital. Now in the process of being patented, the tool will be tried next year in simulated situations with medical students, residents, and attending physicians at Emory before a pilot project begins with patients.

It’s not meant to replace the surgeon's expertise, says Sweeney, but rather to provide more evidence to help surgeons make decisions quickly and accurately.

Emory GI surgeons have been following the research closely, and awareness is already having a positive effect. Reporting on a sample of 2,400 patients, Sweeney estimates that 70 to 100 complications were prevented, based on the number anticipated in such a patient group by the American College of Surgeons National Surgery Quality Improvement Program. With complication costs running anywhere from $5,000 to $50,000 per patient, this represents a savings of $1.5 to $2 million for patients, the hospital, and payers (twice that if the findings are extrapolated to all surgery patients last year).

"HSR is improving our patients' lives,” says Sweeney, "and it’s changing our culture, including how our medical students and residents are learning to think about the care they provide and the system in which they provide it."