PNNL applies risk assessment techniques to health care

Hospitals are for healing. With the help of tools used in the nuclear, aerospace and chemical industries, researchers at Pacific Northwest National Laboratory are helping make sure they stay that way.

In response to a recent standard from the Joint Commission on Accreditation of Healthcare Organizations (JCAHO), more than 18,000 accredited health care organizations in the U.S. and Canada are performing analyses and developing ways to prevent errors that could threaten patient safety. Groups affected by the new standard include not only hospitals, but also clinical laboratories, behavioral health centers, home care facilities and other accredited health care organizations.

Analysts at PNNL are assisting local hospitals in identifying and evaluating situations that may affect patient safety. "We identify frequency and consequence of failures in such a way that we can identify improvements to the system," said Jon Young, program manager for the project. Using risk assessment techniques to evaluate safety is not new to PNNL analysts, who have employed these techniques successfully in the nuclear, aerospace and chemical industries.

Risk assessment related to health care begins with looking at processes and data. "We're interested in finding the underlying problems that can cause mistakes," Young said. "These may be institutional or environmental and may cause other problems as well." Factors could include staff experience level, treatment protocols or interdepartmental communication.

Hospitals call underlying problems "latent failures," things that exist all the time, but don't become visible until something happens that acts as a catalyst. That's when patient safety may be threatened.

PNNL analysts are using risk analysis techniques to analyze existing processes related to patient safety. For example, hospitals already assess some patient safety issues, such as slips and falls, devising a management plan to deal with patient vulnerability. As part of the program, hospital staff might place a placard on the patient's door advising of high risk. "But what happens if the patient is taken to radiology and no one can see the placard?" Young asks. "This is an example of the kinds of things we're starting to find."

In another approach to patient safety, the Laboratory is applying powerful techniques to analyze data in a pilot study with U.S. Pharmacopia, an organization that collects data on medication errors from hospitals nationwide. "We're using our data mining and data visualization technologies to analyze medication error data and put them into a risk context," Young said. The pilot study is currently being funded by Battelle, which operates PNNL for the U.S. Department of Energy, and will expand if grant funding comes through. "This is an exciting opportunity to refine the understanding of errors because we'll have hundreds of thousands of records on medication errors," Young said. Eventually, study results may be published in a report and go back to member hospitals.

"We have other powerful risk analysis technologies we hope to apply in the future. This is just a starting point," Young said of PNNL's work with risk assessment at hospitals. Ultimately, analysts hope to apply risk analysis and data visualization techniques to help hospitals rank problems and determine which issues they'll address first and how.