Hospital Computer Systems Failing to Deliver
One study finds they facilitated medical errors
TUESDAY, March 8, 2005 (HealthDay News) -- Reports over the past few years of increasing numbers of patient injuries and deaths due to medical errors sent hospital administrators scrambling for computerized solutions.
But two new studies suggest that, in many cases, these high-tech systems have left doctors and nurses increasingly frustrated while providing little evidence of real benefit to patients. In fact, one widely used system actually helped foster medication errors, researchers found.
The reports appear in the March 9 issue of the Journal of the American Medical Association.
"Based on our current best evidence, I don't think you can say that these systems will necessarily improve health in as meaningful a way as the vendors of these systems are claiming," said Dr. Amit Garg, a professor of medicine and epidemiology at the University of Western Ontario, and lead author of one of the studies.
Garg's team conducted a review of all the literature to date on "computerized clinical decision support" (CCDS) systems, which are designed to help health-care workers make better choices in daily patient care.
Looking over 100 studies published between 1998 and September 2004, they found that many CCDS systems reviewed did improve "practitioner performance" -- especially those systems designed to help remind patients to get regular preventive screenings (i.e., mammograms, cholesterol check-ups), diagnostic systems, or software that helped doctors prescribe or dose medications.
However, Garg added one important caveat to those findings: Computer systems were much more likely to be awarded high marks for performance if the people who conducted the study had also helped design the system, as happens in many large health-care centers. In studies without that potential bias, in-house computer networks tended to get somewhat poorer ratings, he said.
And while computers may have helped improve certain health-care processes, there's still no evidence that any of this translates to significantly improved patient outcomes, Garg said.
"Even the newer studies aren't showing improvements there," he said.
A more ominous note was sounded in the second study, which focused on one hospital's two-year experience with a widely used "computerized physician order entry" (CPOE) system. According to researchers at the University of Pennsylvania School of Medicine, that in-house network often helped promote medical errors.
In the study, the researchers worked closely with 261 health-care staffers at a large, 750-bed hospital between 2002 and 2004. They observed doctors', nurses' and other workers' daily use of the hospital-wide CPOE system, and conducted frequent one-on-one interviews and focus groups to get staff feedback on the technology.
The researchers concluded the system "often facilitated medication error risks, with many reported to occur frequently."
Doctors, nurses and office staff were often frustrated, the Pennsylvania team found, with frequent system crashes and maintenance shutdowns that could take hours. And computer-based communication between pharmacies and the hospital was often problematic and helped lead to inaccurate dosing of patients or improper and potentially harmful gaps in drug treatment, the study found.
Also, since patients' names were grouped alphabetically (as opposed to by room number), patients with similar names were easily confused as workers quickly scrolled through lists and screens, the study found. Nurses complained they often had to interrupt patient care to find a working computer and then laboriously log in complicated medication information. In some cases, staffers said they needed to go through up to 20 screens to find a patient's full medication record, according to the report.
The bottom line: While it's unclear how many errors the new technology reduced or eliminated, the researchers said they counted 22 different ways in which it actually "facilitated" potentially serious medical mistakes.
Not everyone agrees these two studies do justice to these emerging technologies, however.
"First of all, [the Pennsylvania team] didn't evaluate whether or not computer entry reduces medication error rates, and there's very strong data that CPOE does reduce errors," said Dr. David Bates, chief of the division of general internal medicine at Brigham & Women's Hospital in Boston and a long-time proponent of these types of systems.
He also believes the CPOE system used in the Pennsylvania study is already outdated.
"If they would've evaluated one of the newer applications, there would've been much fewer problems -- many of the things they mentioned would simply not have been an issue," Bates said. "There's no way, now, that you'd have to go through 20 screens to look at a summary of medications. Our software here at Brigham doesn't have many of these problems, largely because we've worked on things."
Everyone agreed that the computerization of health care is here to stay, and that it would be foolish to discard the idea because it has not yet reached perfection. Still, Garg believes health-care workers and hospital administrators may be losing patience -- and money -- with some of these new technologies.
"Computer systems are still in the testing phase," he said, "they are still not really 'ready for prime time.' These technologies need to be tested rigorously, because things that on the surface we think might be beneficial can have unforeseen adverse outcomes."
And he pointed out that a flawed computer system can still cost a hospital millions of dollars.
"At the end of the day, what if we were to spend that money in ways that would better improve the care and health of our patients?" he asked. "That's why this review is so important -- you want to make sure the net benefits outweigh the incremental costs."
To learn how patients can play a role in preventing medical errors, visit the Agency for Healthcare Research and Quality.