Learning from medical mistakes
Although medical errors don't always result in physical harm, they're events physicians and other clinicians aim to avoid. To assist in designing a system that could help them learn from errors, doctors from the Wisconsin Academy of Family Physicians and nurses and medical assistants in family practice recently concluded a series of focus group sessions. In these sessions, they voiced their opinions about the requirements for a user-centered error-reporting system. The centralized system would supply medical staff with real-time feedback and a searchable database of lessons learned.
From transcripts of the sessions, an astounding 82 themes emerged, says Industrial and Systems Engineering Assistant Professor Ben-Tzion Karsh, who conducted the study with Emeritus Family Medicine Professor John W. Beasley and industrial engineering graduate student Kamisha Hamilton. "In order for people to want to use this kind of a system, the main purpose would have to be for system improvement," he says.
Focus-group participants discussed everything from who would "own" or manage the system, what kind of data would be reported, and the system's potential punitive implications to how to motivate staff to use it and how it should be integrated into daily work.
Some of their conclusions were unexpected, says Karsh. While the researchers suspected that doctors, nurses and medical assistants wouldn't want a system through which they might be held criminally accountable, they would like the public to know it existed. "So medical staff weren't trying to abdicate responsibility for problems," he explains. "They felt very strongly that the public should be aware of the way the system works so they could understand it was to improve the quality of care."
And while study participants believed such a system should exist, another interesting revelation was that medical staff had varying levels of comfort about what data — such as the type of nurse, type of doctor, patient age, level of English proficiency — to report. "For this to work, we couldn't have many required fields. We would have to leave it, at least in the beginning stages, to allow the people who are reporting to report as much information as they felt comfortable reporting," says Karsh.
In addition, participants stated strongly that if the system weren't integrated into users' current work flow, then it wouldn't work. Nurses especially were concerned about how long it would take to submit a report, says Hamilton. "Any type of lengthy, drawn-out reporting procedure with tons of paperwork, make a phone call, do whatever else — there's just no time for it. I think their main point was that usually an event would happen when they were extremely busy," she says. "If you've got to take 20 minutes aside when you're already busy, it's not going to happen."
Like their feelings about what data to report, each potential user had a distinct level of comfort with the medium — computer, paper or phone — they might use to submit a report. "We know from good engineering design that if you want people to be able to access any type of system, you need to have flexible design," says Karsh. "And so another principle that came out of this was that we were going to need to develop something that has flexible design and flexible interface."
Equally as important were the benefits medical personnel would derive from the system. Again, focus-group participants said they wouldn't be likely to use a system that didn't generate a nearly immediate result from their reporting. "People were in favor, for example, of weekly E-mail alerts," says Karsh. The E-mails might highlight a specific event, explain how it happened and then offer a solution or preventive measure.
A common thread throughout the focus-group sessions, trust emerged most notably when participants talked about what entity would own and manage the system. "What came out was that people were very uncomfortable with it being in a state agency, because politics could get involved," says Karsh. Some doctors, nurses and clinicians saw a university as a likely manager since both groups could learn from the data.
Conversely, however, other medical staff regard universities with hospitals and health plans as competition, so the group concluded a non-governmental agency might be best-suited for the task. A multidisciplinary team that includes risk managers, pharmacists, nurses, doctors, physician's assistants, engineers and others would analyze the data. "I think we learned a lot about how much work is really going to be involved to make something like this successful," says Karsh.
Although the researchers directed their study only at family-practice physicians and staff, it might be a springboard to learning what kind of system would suit staff in other specialties, he says.
The three currently are validating their data and writing summaries and reports they will submit to legislators, lobbyists and medical associations, among others. A grant from the Industrial and Economic Development Research Program at UW-Madison and an in-kind contribution from the Wisconsin Academy of Family Physicians funded the study.