"We examined a particular type of pain pump and estimated how many deaths are likely to occur as a result of programming errors," says Professor Kim Vicente of U of T's mechanical and industrial engineering department. "The number was astounding."
The researchers multiplied the number of reported deaths connected to pain pump use (between five and eight, according to a search of the U.S. Food and Drug Administration database) by a "general medical under-reporting factor," as identified by previous researchers. "Roughly, for every one death you hear about in general, there are between 13 and 83 that you don't hear about," he says. "By these estimates, the number of deaths from programming errors with this pump could be as high as 667."
In the April issue of the Canadian Journal of Anesthesia, Vicente and co-authors Karima Kada-Bekhaled, Gillian Hillel and Andrea Cassano of U of T's cognitive engineering lab and Beverley Orser of anesthesia at U of T and Sunnybrook and Women's College Health Sciences Centre say the health system needs to keep better track of problems that occur with medical equipment. Previous epidemiological studies have shown that accidental death or adverse consequences in general are vastly under-reported to health authorities. Reporting rates range from 1.2 per cent to 7.7 per cent. "When people do things incorrectly and know they're going to be punished, blamed or fired, they tend not to report things," says Vicente. "That's a consistent and well-documented fact about health care; I would estimate that about 95 per cent of the errors that happen don't get reported."
In the case of the pain pump, researchers recount one case that was reported. In 2000, a 19-year-old woman underwent a Cesarean section to deliver a healthy infant. To control post-operative discomfort, the mother was given a patient-controlled analgesic (PCA) pump. PCAs inject morphine or other analgesics into patients via a pre-set program. Nearly eight hours later, the patient was dead from an overdose of morphine due to a programming error on the machine.
The PCA used in this case - an Abbott Lifecare 4100 Plus II Infusion Pump, one of the most commonly used pain pumps in North America - has also been involved in the care of up to seven more patients who died, say the researchers. The problem, says Vicente, stems from the pump's design, which makes it prone to human error when it is being programmed.
This research builds on earlier work by Vicente and other colleagues at U of T. In 1998, they compared programming of the Abbott Lifecare 4100 and one with a modified design. Using computer simulations, they found concentration programming errors were reduced to zero using the modified pump.
To ensure greater patient safety, the researchers recommend that the control panel used to set medication levels for PCA pumps be redesigned using human factors engineering techniques. Vicente is a leading expert in this growing field which examines how people think about, interact with and use technology.
Medical equipment, by its very nature, is complicated, says Vicente, this year's Jerome Clarke Hunsaker Distinguished Visiting Professor at the Massachusetts Institute of Technology's Department of Aeronautics & Astronautics. Manufacturers who design devices that take into account human factors principles will have devices that are easier to use and less prone to human error. In the case of medical devices like pain pumps, this may result in fewer accidental deaths and adverse events, he says.
"When it comes to devices like medical equipment, we always need to keep in mind that it's people who are ultimately using and programming these things."
This research was supported by the Natural Sciences and Engineering Research Council of Canada.
CONTACT:
Janet Wong
U of T public affairs
416-978-5949
jf.wong@utoronto.ca
Kim Vicente
Department of Mechanical and Industrial Engineering
617-253-5624
kjv@mit.edu
Journal
Canadian Journal of Anesthesia/Journal canadien d anesthésie