Hospitals: Boosting Medical Safety Innovation

Photo: Data sheet

Hospital discharge data sources can be a useful source of information for patient safety; ©
Yukchong Kwan

As health care personnel race to improve the quality of their care to save lives and prevent unneeded harm, a new study indicates there is more they can do to learn about what errors are occurring and why.

Researchers from the Drexel University School of Public Health demonstrated a systematic analysis of hospital administrative data for patient safety at a population level. They say that health care organizations have an untapped opportunity to use their own administrative data in this way as a "springboard to problem identification" at the leading edge of preventing even those medical errors that are not yet preventable. "For example, a patient may receive a drug in the Emergency Department and develop an allergic reaction, but did not have any known allergies at the time of treatment," said Doctor Jennifer Taylor, an associate professor at Drexel who led the study. "While such events may not be deemed to be preventable now, we need to start tracking them so our research and development colleagues know what's next in the prevention pipeline."

Based on this premise, that preventing medical errors requires a good understanding of when and where all such errors, or so-called "patient safety events" occur – including those that are nonfatal, those that are not yet preventable and even those that do not cause noticeable harm – Taylor and colleagues analyzed large-scale data on hospital stays (recorded in discharge data) in Pennsylvania during one year. They compared hospital stays with and without patient safety events, to describe patterns, demographics and differences associated with such events. They found that nine percent of the hospital discharges in Pennsylvania in 2006 were for stays with a patient safety event. On average, patients who experienced an adverse event were older, white and male; patient safety events added an average of 35,000 dollars to the cost, and 3 days to the length, of a hospital stay.

"The percentage of discharges that had a patient safety event is in the range of other studies we have seen in the U.S. and around the world," said Taylor. "While this figure may be a bit startling, it is not a cause for alarm, in that many of the events that we found are adverse events for which there are no known prevention strategies," as in the example of the unexpected allergic reaction.

For their analysis, Taylor and co-authors used standard International Classification of Disease (ICD-9-CM) diagnosis codes that indicate each patient's primary and secondary diagnoses and causes of injury in hospital discharge data. These discharge data are typically used for billing, but are also used for some public health surveillance efforts such as reporting on communicable disease. The ICD-9-CM codes provided a population-level picture of many preventable and non-preventable adverse events. The researchers combined this data source with an algorithmic patient safety indicator developed by the Agency for Healthcare Research and Quality (AHRQ), which has been demonstrated as a useful indicator of specific adverse events for which known prevention strategies are available.

Taylor and colleagues point out that hospital discharge data sources such as these are easily accessible to administrators and safety personnel and can be a useful source of information for patient safety. By describing patient safety events at the population level, their method capitalizes on the consistent standardized information the discharge data provide, and could be used by any health care entity across the U.S. and the world that utilizes the ICD system for coding.; Source: Drexel University