Information technology can help detect adverse events

Electronic screening of hospital discharge records for adverse events could provide a reliable way to detect medical errors, according to a study in the Journal of the American Medical Informatics Association.

"Although simple keyword searches are not specific enough to be helpful, the use of more sophisticated searches could eliminate the need for hospitals to manually review discharge reports," said Dr. David Bates of Brigham and Women's Hospital in Boston.

Bates and colleagues at Brigham and Women's developed a computerized screening tool designed to help eliminate the need for hospital panels to manually review discharge charts. The tool scans for words or phrases, such as "shortness of breath" or "transfer," that could indicate adverse events.

The study reports on 424 randomly selected discharge records from January to July of last year. Nearly 45% percent of the discharge summaries contained keywords that could indicate adverse events. After physicians reviewed the records, the tool was found to have actually detected 131 adverse events.

Researchers concluded that although the screening tool performed moderately well, its lack of targeted search terms caused too many false-positive hits, according to the study.

Using more specific keyword queries that incorporate natural language processing - programs that can scan a free text narrative and understand languages that humans use - could improve the search tool so that clinicians don't have to manually review records, Bates said.

"Although patient safety is a major problem, most hospitals rely on spontaneous reporting, which detects only a small minority of adverse events," he said. "As a result, problems with safety have remained hidden."

Chart reviews can detect adverse events in research settings, but they are too expensive for routine use. Information technology techniques can detect some adverse events in a timely and cost-effective way, in some cases early enough to prevent patient harm, Bates said.

The study concluded that electronically screening text in medical discharge reports was a promising way to detect many types of adverse events, particularly as more medical information becomes computerized.