Short-term return visits to the emergency department (ED) are increasingly used as a measure of hospital performance. The theory is that people who return for treatment and admission must have received poor quality care or were improperly discharged the first time.

A new study by the University of Washington Schools of Medicine and Public Health and collaborating institutions found just the opposite.

“The findings were totally contrary to our hypothesis,” said lead author Dr. Amber Sabbatini, a University of Washington emergency-medicine physician and acting instructor in the Division of Emergency Medicine. “We had expected their outcomes to be worse than patients who were admitted on the initial ED visit, treated and released [in that same time frame]. But in fact, ED return admissions led to lower in-hospital mortality, with less time in the ICU and lower costs.”

Dr. Sabbatini and colleagues studied data from more than 9 million patients who visited 424 hospitals in Florida and New York in 2013. Patients with an initial visit were identified and followed for revisits within seven, 14 and 30 days. The study was published in the Journal of the American Medical Association.

“It raises an important question as to whether repeat ED visit should be deemed as medical errors committed by not admitting the patients to the hospital during their index ED visit,” said co-author Dr. Anirban Basu, professor of health services at the UW School of Public Health and a professor of pharmacy and adjunct professor of economics. “The results show that we should be careful in making such an overall judgment and perhaps try to understand and identify specific conditions where repeat ED visits are a consequence of inadequate care rather than being a part of the continuum of care for these patients.”

To do this correctly, Dr. Basu added, researchers need detailed electronic health data beyond just those available in claims databases. “Such data are hard to come by and, even when available, are extremely expensive for researchers to access,” he said.

Participating institutions were the University of Michigan and the University of California, San Francisco.