One-third of U.S. hospital admissions can expect a medical error, an infection of other adverse outcome, U.S. researchers said.

"Without doubt, we've seen improvements in healthcare over the past decade, and even pockets of excellence, but overall progress has been agonizingly slow," Susan Dentzer, editor in chief of the journal Health Affairs, says in a statement.

David C. Classen of the University of Utah and co-authors at the Institute for Healthcare Improvement compared three methods for detecting adverse events in hospitalized patients -- the institute's own Global Trigger Tool and two methods most commonly used to track patient safety.

Among the 795 patient records reviewed, voluntary reporting detected four adverse events, the federal government's Agency for Healthcare Research and Quality Indicators detected 35 and the Global Trigger Tool detected 354 events -- 10 times more than the AHRQ method.

In other words, the AHRQ indicators of the federal government and voluntary reporting by hospitals missed more than 90 percent of adverse events identified by the Global Trigger Tool, Classen says.

The researchers say their findings are conservative, because their study is based on review of medical records and real-time observation in hospitals, which would most likely detect more adverse events.