STUDIES SHOW THAT SOME MEDICAL DEVICES ARE TOO COMPLEX
COLUMBUS, Ohio -- Sometimes new, high-tech
medical devices are too complicated for the doctors and nurses who use
them, according to a pair of Ohio State University studies.
The studies revealed that poor human-computer
interfaces can create errors in the set up and use of these devices.
The studies also showed how medical professionals work around these difficulties
in order to avoid making mistakes that would endanger patients.
Computer technology has enabled engineers to design medical devices
that perform many tasks, and hospitals save money when they buy such devices.
Too often, though, the cumbersome computer interfaces in these devices
hamper the efforts of people who use them.
The studies appeared in a recent issue of
the journal Human Factors, where David Woods, professor of industrial and
systems engineering and anesthesiology, wrote, "There is a real risk of
error when designers create systems that possess every
conceivable option, multiple modes, and cryptic displays."
In the first study, Woods and his colleagues observed how anesthesiologists
interacted with a new anesthesia monitor in the operating room. In
the second, they interviewed home nurses who supervised pregnant women’s
use of infusion pumps that dispense medication to prevent premature labor.
In both studies, the researchers saw the same
problems over and over again, like the need for medical professionals to
remember complicated sequences of commands. Also, the anesthesia
monitor and the infusion pump operated in multiple modes, without clearly
displaying which mode they were in at any given time.
"New technology can reduce error, but it can
also create new opportunities for error," Woods said. "And sometimes the
new errors, though infrequent, may contribute to more catastrophic consequences."
Woods stressed that the kind of errors he’s
talking about don’t result from a malfunction in the medical device.
"It’s not that the device is unreliable, but
rather that the design of the device does not take into account how people
need to interact with it. The problem is not a breakdown in the engineering
of the device, it’s a human-computer breakdown," Woods said.
In both studies, no patients came to harm, but the medical professionals
did have to tailor the devices to meet their needs and to avoid mishaps.
In the study of the anesthesiology monitor,
for example, technicians experimented with the monitor outside the operating
room until they figured out how to call up the most useful displays and
how to configure multiple windows on the screen. However, when the physicians
called up one window during operations, the computer would automatically
change to other windows, obscuring important patient information on the
computer screen. Users had to discover when the undesired window configuration
occurred and develop new procedures and reminders to "undo" the computer’s
reconfiguration.
Because the user interface to the home infusion
pump was complex and commands were hard to remember, nurses changed procedures
or created new procedures to help the patients avoid or work around trouble.
For instance, patients were supposed to change the syringe on the infusion
pump whenever an alarm signaled the syringe was nearly empty, which happened
in the middle of the night.
Even for people who are fully awake, the clumsy
computer interface makes this a relatively complex task. So nurses
told the women to change the syringe at the same time every day even if
the syringe wasn’t empty so that they wouldn’t have to fumble with the
pump and interface in the middle of the night.
In both studies, the tailoring that the users
had developed worked. But Woods said that the fact users had to tailor
the devices at all indicates that the technology isn’t as user-friendly
as it should be.
And one problem remains: because the devices
don’t give good feedback, users still can have difficulty detecting errors
in the future and correcting them in time.
Woods said that while error prevention is
important, so is error correction. "Safety comes from recognizing
that while we try to prevent errors, we also need medical devices to give
us time to detect and recover from errors before we get bad consequences,"
Woods said.
Woods and his colleagues hope to synthesize
these and other results to show medical manufacturers and regulators how
poor human-computer interaction in the design of infusion devices can lead
to user errors. The goal is to help people understand the importance
of human-computer interaction in effective medical products and to begin
to apply concepts and techniques from Human Factors.
Also, the American Medical Association is
creating a National Patient Safety Foundation, and Woods is helping the
foundation set up its research program. One of his goals is to help
people develop user-centered medical systems to improve patient safety.
"If we understand the sources of errors, we can find ways to
make good devices even better."
Contact: David D. Woods, (614) 292-1700; Woods.2@osu.edu
Written by Pam Frost, (614) 292-9475; Frost.18@osu.edu
Return
to the current month abstract page
Return
to the Research page
Return to the OSU Homepage
Go
to the Reasearch, Newsfeature, and Cancer Report Archive