Study: CPOE facilitates 22 types of medication errors
Study: CPOE facilitates 22 types of medication errors
Systems must be implemented carefully
Although computerized physician order entry (CPOE) is expected to significantly reduce medication errors, systems must be implemented thoughtfully to avoid facilitating certain types of errors, according to a study published in the March 9 issue of the Journal of the American Medical Association (JAMA).
The study was supported by the Agency for Healthcare Research and Quality (AHRQ).
The study, Role of Computerized Physician Order Entry Systems in Facilitating Medication Errors,1 examined the use of a CPOE system at a tertiary-care teaching hospital from 2002 to 2004. It is based on interviews with medical staff, focus groups, shadowing staff as they worked, and a survey of interns and residents.
The study identified 22 situations in which the CPOE system increased the probability of medication errors.
According to the authors, these situations fell into two categories:
- information errors generated by fragmentation of data and hospitals’ many information systems (45% of the errors);
- interface problems between humans and machines, where the computer’s requirements are different from the way clinical work is organized.
Three-quarters of the house staff reported observing each of the error risks, indicating that they occur weekly or more often.
According to the study, flaws include the following:
- Medical staff may look to the CPOE system to determine minimal effective or usual dosage for infrequently used medications. However, the CPOE system may only reflect dosage sizes available at the pharmacy, which may differ from the minimal or usual dosage that should be prescribed.
- Clinicians might select the wrong patient file because patients’ names and the names of drugs can be hard to read, computer mice often are imprecise, and patients’ names do not appear on all screens.
- A patient’s medication information is seldom synthesized on a single screen. Up to 20 screens might be needed to see all of a patient’s medications, increasing the likelihood of selecting a wrong medication.
- Because of the patient load and multiple tasks, nurses may be unable to enter timely information on the computer about the administration of drugs. The delayed information may affect later medication and clinical decisions.
- Computer downtime, whether for maintenance or in the event of crashes, can result in delays in medications reaching patients.
Why errors occur
Ross Koppel, PhD, sociologist at the Center for Clinical Epidemiology and Biostatistics in the University of Pennsylvania School of Medicine in Philadelphia, and lead author of the study, offers several insights into why the errors occur.
"Fragmentation occurs because some software is built up over time that has never been integrated into the way the hospital actually does its work," he explains.
"Several systems can be in use, and there can be a lack of thoughtfulness about the flow of information. Why, for example, does it take 20 screens to see all of the patients’ medical records? Why are screens for ordering meds different than those for reordering?" Koppel asks.
To minimize these types of errors, he continues, you need to look at the system as a whole.
"When you are negotiating with a vendor, you should think of the total flow of products, services and people, and make sure the system is responsive to the way they work," Koppel advises. "Software can’t force doctors, nurses, and others to twist like pretzels around it; instead, it should facilitate their work."
That attitude should continue once the system is in, he emphasizes. "Software was created by humans for humans. When you go to the IT people and say something is not working well, and they roll their eyes and say it would require them to change a lot of lines of code, the proper response of the hospital should be, Yeah, so?’"
As for difficulties with humans and computers interfacing, while some may be due to "really sloppy programming," they also can be the result of not paying attention to the display, or to basic ergonomics, Koppel adds.
"This can include having terminals in really busy places where there are thousands of distractions, or where the lighting is miserable, or the noise level extraordinary," he notes.
Problems also can arise if attention is not paid to the way in which the work itself is done, Koppel continues.
"If the first thing I do is look up past history, and the second is lab tests, the third ordering new drugs, then that should be a seamless interface," he adds.
Koppel posits a hypothetical situation in which a nurse is trying to administer a group of nine meds.
"Let’s say she’s given four, and a patient throws up or is called away to radiology. What happens is, if the nurse has to go through many, many screens, she may not be able to administer all the meds. Or if she’s in a hurry, she may have to choose between correcting a chart contemporaneously or actually give the med. So the nurse chooses to give the meds in a timely manner, but charting is delayed," he says.
That delay has implications for time-sensitive drugs such as insulin. "Docs then have to make critical decisions based on when they think the drug was given," Koppel notes.
Some of these problems could be solved by putting mobile input devices on the meds cart, he suggests. "The cost is trivial compared to the operating budget of the hospital."
New systems still have problems
Some critics of the study have noted that the system used at the time was an older version no longer used, and newer systems may eliminate some of the problems Koppel has observed.
However, he does not entirely agree. "Whenever you change something, there will be cascading implications, which must be carefully examined and addressed," Koppel insists. "It’s a constant game of catch-up."
One of the newer software programs requires physicians to fill in some fields on a screen before they can proceed to the next screen, he explains.
"Some docs may say, This has no relevance to me,’ and start making up information because they are in a hurry.
"They know the made-up information is benign in terms what they ordered, but someone may come in later and make an error based on that information. So with this system, benign information can come back to bite you," Koppel points out.
To minimize such problems, he advocates five recommendations, as outlined in the JAMA article:
- Focus primarily on the organization of work, not on technology.
- Aggressively examine the technology in use.
- Aggressively fix technology when it is shown to be counterproductive.
- Pursue errors’ "second stories" and multiple causations to surmount the barriers enhanced by episodic and incomplete error reporting.
- Plan for continuous revisions and quality improvement, recognizing that all changes generate new error risks.
For quality managers in particular, Koppel recommends eternal vigilance.
"You must focus on every aspect of the system; get down on the floor with it. Virtually everything we have been talking about is quality management. You’ve got to go down on the floor, analyze what’s going on and how the system is used. You have to, in essence, play minisociologist and find out why people use the system the way they do," he adds.
Reference
1. Koppel R, Metlay JP, Cohen A, et al. Role of computerized physician order entry systems in facilitating medication errors. JAMA. 2005; 293:1,197-1,203.
Need More Information?
For more information, contact:
- Ross Koppel, PhD, Sociologist, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania School of Medicine, Philadelphia. Phone: (215) 576-8221. E-mail: [email protected]]
Subscribe Now for Access
You have reached your article limit for the month. We hope you found our articles both enjoyable and insightful. For information on new subscriptions, product trials, alternative billing arrangements or group and site discounts please call 800-688-2421. We look forward to having you as a long-term member of the Relias Media community.