By shifting from paper-based or verbal orders to electronic ordering, advocates say, hospitals can eliminate errors caused by illegible handwriting or ambiguous voice messages. In addition, CPOE systems include decision support functions that steer physicians to making appropriate decisions based on medical evidence. And they provide alerts that warn physicians about orders for drugs or procedures that have the potential to harm their patients.
But so far, only about 8% of U.S. hospitals have implemented CPOE, the Leapfrog Group estimates. And a recent study by the Washington-based employer consortium found that hospitals that have adopted the technology are facing some serious challenges.
Some 100 hospitals recently used Leapfrog's CPOE Evaluation Tool, which offers scenarios for testing certain order entry functions. And the results were disappointing, says Leah Binder, Leapfrog's CEO.
For example, although the vast majority of drug allergies were caught by the systems, some were not. Also, some hospitals' CPOE systems failed to offer an alert that a drug should be taken with food. And a few hospitals' systems even failed to prevent a potentially fatal medication order in the simulation.
"It's extremely complicated to set up these systems properly," Binder says. "They have to be updated continually. There are always going to be bugs that hospitals need to address on an ongoing basis."
Hospitals face the challenge of customizing the off-the-shelf software to meet their specific needs, Binder adds. "These are not plug-and-play systems," she stresses.
Many experts also warn that hospitals must take extraordinary steps to avoid "alert fatigue" caused by vendors' systems that are pre-programmed to display so many alerts that physicians begin to ignore most of them. Moreover, successful deployment of CPOE requires time-consuming re-engineering of care processes, they stress.
The key to using CPOE to consistently prevent medical errors, Binder argues, is to continually test the technology and refine it. All 33 hospitals that were top scorers in the Leapfrog Group's annual quality survey used the CPOE evaluation tool to help them qualify for the recognition (see sidebar, page 22).
"Many institutions have no idea of how they are doing on providing decision support," says David Bates, M.D., chief of the general medicine division at Brigham and Women's Hospital, Boston. "The CPOE evaluation tool helps leadership to measure where they are with decision support."
Refining The Tool
Bates helped Leapfrog Group refine the tool, providing feedback that the initial version was far too difficult to use and did not focus on the most common orders that affect patient safety. By using the revamped tool, hospitals can more precisely target their efforts to refine decision support in their CPOE systems, the physician says.
"We have a long list of things that we want to add that we haven't gotten around to yet," he says. "The tool helped us refocus our efforts."
But another physician whose organization used the tool says the test focused far too heavily on the triggering of alerts and not enough on testing the logic embedded in decision support.
"We need to test the actual functioning of the order sets," says Charles Ross, M.D., chief medical information officer for Summa Health System, which owns six hospitals in Ohio. The key to a successful CPOE deployment, Ross stresses, is building logic into the order sets that, rather than triggering an alert, steers the doctor to the right decision.
For example, a well-designed CPOE system would prevent doctors from ordering certain drugs to treat pneumonia, rather than displaying various alerts about the inappropriate drugs that could be ignored.
"If you met all the requirements of the Leapfrog tool, it would lead to some over-alerting of physicians," contends Lori Idemeto, pharmacy informatics specialist at Virginia Mason Medical Center, Seattle.
She observed, however, that the tool helped the organization identify some areas for improvement. For example, the hospital tweaked its order sets to provide more guidance on single and cumulative drug dose limits.
Evidence Is Key
Physicians will embrace CPOE and clinical decision support if they know the technology is grounded on evidence-based medicine that considers local needs, says Michael Zia, M.D., chief medical officer at Decatur (Ill.) Memorial Hospital, a 350-bed community hospital. "One of the biggest issues physicians have with evidence-based medicine is that it may be the best way to treat patients at Johns Hopkins but not necessarily at Decatur Memorial," says Zia, who also serves as the hospital's vice president of quality management. For example, patients in central Illinois may have a history of resistance to antibiotics that's radically different than patients on the East Coast, he notes. "We'll use data to define evidence-based medicine to meet our needs here."
Hospitals cannot possibly know how to optimize their CPOE systems until they get some experience using the technology, Zia argues. Decatur Memorial, for example, initially gave physicians two options for ordering blood transfusions using a CPOE system from McKesson Corp., San Francisco. One option involved using decision support to justify the order; the other involved simply typing in an "a la carte" order. "We found that physicians were looking at evidence-based medicine criteria using the decision support function only 60% of the time," he says. "So we decided to hide the a la carte order function from the CPOE menu."
Alert Fatigue
As early CPOE adopters fine-tune their systems, they're often trimming the number of alerts the applications display to physicians placing orders for medications and tests.
"When you start throwing out alerts for everything, you run into alert fatigue," says Ross of Summa Health, which uses CPOE software from Eclipsys Corp., Atlanta. "Residents and physicians blow right by the alerts, and CPOE becomes a negative learning tool."
Be the first to comment on this post using the section below.