Error reporting: Even if it doesn’t seem broken, it still might need fixing
Medical error reporting systems get an overhaul
Just because your hospital doesn’t report medication errors in large numbers doesn’t mean you don’t have a problem. Indeed, according to Sharon Lau, a consultant with Medical Manage-ment Planning in Los Angeles, members of her group of children’s hospitals that share data contend that a low error rate really means your staff are underreporting the problem.
Since the Institute of Medicine released its report last November detailing the problem of medical errors, the media, the public, and many hospitals have focused on the issue. For many facilities, the problem has become one of overhauling their systems, creating new ways to report and catch errors, and encouraging all staff to participate in the process.
Lau’s group started looking at the issue back in 1993, and for six years struggled repeatedly with the question of how to make sure that the data the group shared were reliable and accurate. A big problem was defining what was a medication error. The group met in Dallas in January to hash out definitions. (See list of definitions, p. 77.)
Coming up with a definition solves the issue of "what is the numerator," Lau explains. But the group had similar problems — and continues to have them — on determining the definition of the denominator. "[Our] definition has always been for dispensed doses," she says. "However, in some instances, hospitals have been unable to retrieve that information from their information systems. They have, therefore, submitted administered/ billed doses as their denominator."
She adds that the group has discovered a correlation between medication errors per dose and medication errors per day. That means "when the error rate per dose is high for one hospital, its error rate per patient day also tends to be high. Conversely, when the rate for a hospital is low for errors per dose, its rate of errors per patient day also tends to be low."
This does not necessarily mean that patient days is a more accurate denominator than doses even when adjusted for case mix index. "It may simply be an interesting statistical observation."
And even if the hospitals agree on how to report the data, there is a continuing question about how many errors are actually reported.
They all do it their way
Roger Resar, MD, is the "change agent" at Luther Midelfort Hospital in Eau Claire, WI. The 300-bed facility is part of the Mayo Health System. Resar doesn’t like to compare his facility’s errors with anyone else’s, in part because systems vary so greatly from hospital to hospital. "I can’t compare to other hospitals until the others are doing what we are," he says. "It’s a moving target. So for now, I compare us to us."
Resar says the problem exists at every hospital and that he believes it exists in similar rates at most hospitals. "But once we found we were like every other hospital — that is, mediocre — we knew we had to make a change. Our mission isn’t to strive for mediocrity."
Steve Meisel, PharmD, assistant pharmacy director at Fairview Southfield Hospital in Edina, MN, also suggests the problems are similar from hospital to hospital. "I think everyone is victimized by the same bad systems," says Meisel. At his 300-bed facility, about 40% of the problems are prescription errors, about 35% are administration errors, and the rest relate to transcription.
However, to build a system that tackles medication errors seriously, says Resar, you have to change your facility’s culture. "You have to have leadership that is convinced that this is an area worth spending resources on, and you have to make it a reportable issue."
Meisel agrees that along with changing systems, implementing more automation, and improving the management of the formulary, changing the culture of a facility is the most important part of solving the medication error problem. "Addressing this problem has to be one of your top five priorities," he says. "Otherwise it isn’t a quest but a hope. And a hope isn’t a plan."
Because no facility observes staff all the time, error reporting is by its nature voluntary, says Meisel. "You have to rely on someone telling you about a mistake that either they made or they witnessed." Because it is voluntary, hospitals face the problem of underreporting. "But we think that we underreport the kinds of errors, when and why they occur equally."
Along with not knowing how many errors a hospital has, there is also a lack of data on near misses, which can be very instructive in developing patient safety programs. But Northern Michigan Hospital in Petoskey, MI, took an idea from the supermarket — the bar code — and used it as part of a patient safety system that should not only reduce errors, but also provide a way of accurately counting errors and near misses.
Communicating, verifying
Using a system from Bridge Medical, a San Diego-based company, that works with bedside computers and radio wave-controlled communications systems that are wired into the ceilings of hospital hallways, Northern Michigan Hospital started a pilot project with 36 beds in late 1998.
Nurses scan the drug to be administered, the patient ID bracelet and their own ID badge. The system then verifies the "five rights" — right patient, right drug, right dose, right time, right route of administration (oral, intravenous, or injection). The system also checks for safe dosing levels and alerts nurses of potential hazards with look-alike, sound-alike medications.
Because of the computers and radio communications the system includes, changes in medications, dosage levels and other patient information can be instantly communicated from hospital information systems to the bedside unit, keeping nurses on the floor up to date on changes.
The new system automatically records when a medication dose is given, which staff member gave the medication, and other information. It produces reports that allow managers to monitor the medications given to patients and helps hospitals identify opportunities for improvements in their medication administration procedures.
"With the system, we have the capability of tracking medication events and determining whether an error was prevented," says Trudy Day, RN, clinical nurse manager at the hospital. "Without the system, we are unable to identify all errors, let alone near misses, since the clinician involved is often unaware that an error has occurred."
One way to make sure that staff understand the importance of error reporting is to show them the numbers, Meisel says. At Fairview, pharmacists, physicians, and nurse leaders identify and prioritize the problems that need to be addressed. Teams are commissioned to collect the data on particular kinds of errors. Various committees, including the quality committee, the safety committee, and the steering committee responsible for addressing the issue of medication errors, are given that data. "Information on every error is collated and included on a statistical report provided to various committees," Meisel continues. "There is nothing more powerful than a run chart in terms of visualizing your improvement process."
When there are successes, they are shared with the hospital staff at large. Teams that work on the problems are also given rewards and awards for getting the job done.
A lot of people talk about how the system of error reporting should not be punitive, but Resar doesn’t think that’s the most important aspect of creating a good system. "That our system isn’t punitive doesn’t get them to report errors. You have to actively praise people for reporting, do something about the problem, and then report back to them what you have done. They have to see the results, not just fill out some form."
Maybe when all hospitals have systems like Luther Midelfort, Resar says, he will start comparing his facility to others. "We have had a system for a long time where each physician has a way of doing something based on his or her training. They want to use that system. But that variability can cause a lot of errors. We had 14 physicians here who had 28 different sliding scales for insulin protocols. And if you are a nurse working with those 14 physicians, you have a greater chance of making a mistake. We really have to rethink a lot of things to get this right. There are many areas that are ripe for improvement."
"There are three elements to creating change," Meisel concludes: "Will, ideas, and execution. There are lots of ideas out there, and you shouldn’t be ashamed to steal shamelessly. There is no mo-nopoly on patient safety."
Meisel thinks that with all the attention medical errors are getting now, will is not an issue. "But execution means doing something. In health care, we like to study things to death. We gather hundreds of data points, go to 87 committees, and modify the plan. But by the time you start, half the team has died of old age." At Fairview, there is a motto: What are you going to do next Tuesday? "When I sit down with a team to coach, lead, or facilitate, that is my question," he says. "If the answer is study, discuss, meet, review, or plan, the meeting isn’t over. There have to be action words about next Tuesday. Even if you do it with just three patients for two days with one doctor, that is the beginning of change. That is actionable stuff. We can’t discuss things to death. Start small, but start."
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content