By Ericka Wilhelms

Using an Observation-Based System to Track and Reduce Medication Errors

IN APRIL 2003, LANCASTER GENERAL HOSPITAL, a 563-bed community teaching hospital in Lancaster, Pennsylvania, and winner of ASHP’s 2005 Best Practices Award in Health-System Pharmacy, began using AU MEDS (Auburn University Medical Error Detection System) from MedAccuracy as part of a VHA study on bar coded bedside medication administration. Using AU MEDS both during and after the hospital’s implementation of Bridge Medical’s MedPoint system, Lancaster General tracked medication errors related to medication administration and used that data to ultimately improve their medication-use process. Tina Suess, Lancaster General’s Bridge system administrator, contends that typical incident reports will not always provide accurate data regarding medication errors, because they record “only the errors that nurses recognize.” Understanding this limitation, Norma Ferdinand, senior vice president of patient care services, recalls the hospital staff “wanted to identify some way that we could truly monitor our medication-error rate, particularly before and after the Bridge implementation,” and AU MEDS was identified as a viable method to meet that goal.

The two-pronged AU MEDS system comprises peer observation and AU MEDS’ proprietary reporting software, developed by Kenneth N. Barker, PhD, RPh, and Elizabeth Allan Flynn, PhD, RPh, both from Auburn University. After contracting with a facility, MedAccuracy works with the hospital’s staff to identify and train pharmacists and nurses to become certified medication observers, or CMOs. The trained CMOs then observe and record nurses’ administration activities during their normal rounds. According to Suess, after the training sessions, the AU MEDS staff observes the CMOs “to make sure that they can be flies on the wall. As an observer, you can’t be obtrusive or interruptive.”

Lancaster General currently has seven trained CMOs performing two unannounced observations per month, resulting in a total of about 200 observed doses per month. An observer will shadow a nurse from the time a medication is dispensed to the time it is administered. The observer then compares the administration activities to the drug orders, and any deviations are recorded as errors using AU MEDS’ proprietary software. Lancaster General categorizes errors in one of eight groups—unordered drug, extra dose, omission, wrong dose, wrong route, wrong time, wrong form, and wrong technique. Breaking the errors down into those groups helps the hospital to use AU MEDS’ graphical reports to “dig deeper and identify trends,” says Ferdinand.

Following data entry, the AU MEDS software creates Microsoft PowerPoint reports that can include all observations to date or more detailed reports that can be tailored to focus on specific time periods, nursing units, shifts, and so on. Standard reports include:

• Is Our Accuracy Rate Increasing?
• Which Nursing Units Have the Highest/Lowest Accuracy Rate?
• What Kinds of Errors Occur? • How Many Errors Are Clinically Significant?
• Which Drugs/Drug Classes/Drug Forms Have the Lowest Accuracy Rates?
• Which Routes of Administration Have the Lowest Accuracy Rates?
• Comparison of Clues to Causes of Errors

At Lancaster General, the reports are presented to the medication management committee, an interdisciplinary team of pharmacy and nursing personnel who determine whether the reported errors are “user issues or larger systems issues,” explains Suess. According to Richard Paoletti, director of pharmacy, “the hard-line data is analyzed by both pharmacists and nurses—those who are closest to the process.” Furthermore, Paoletti continues, “The pharmacy and nursing staff create the fix. Their solutions are practical, because they are the front-line practitioners.”

Results
In Ferdinand’s opinion, the most useful result of the AU MEDS implementation is the way the system has revealed “the true environment in which our nurses are practicing. We now have the ability to ‘see’ how they use the Bridge technology and to catch workarounds.” During observation, it was noted that bar code labels were not being properly placed on insulin syringes. The bar codes were found stuck to computers, clipboards, and name badges. To correct this problem, pharmacy is replacing multi-dose vials with insulin cartridges that have manufacture-applied bar codes, ensuring the correct product is administered and eliminating the potential for mislabeling. With an improved understanding of their nurses’ working environment, the medication management team can then make better-informed decisions that will positively impact the end users of the Bridge system.

In addition to identifying workarounds, the AU MEDS system has helped Lancaster General cut off potential errors at the pass. Paoletti indicates that using the AU MEDS system before the Bridge implementation “was very helpful in identifying some of the systems issues that could result in medication errors at the bedside. Eventually, the bar code system would catch those errors, but we were focused on preventing them from beginning.” Upon reviewing data generated by AU MEDS, Lancaster General’s pharmacy developed new storage methods for their extended-release pharmaceuticals, affixed alert stickers to certain formulations, standardized the IV solutions available on the nursing units, and separated high-risk medications from the rest of the floor stock, so that nurses could better focus on them. The hospital also adjusted its medication administration times, based upon observations of the nurses’ medication passes. Paoletti concludes, “We have made a lot of systems changes based on our observation data. By minimizing—upfront—the errors the bar code system would catch, we created a better process overall.”

In addition, Suess explains that the hospital’s CMOs “can actually identify some variations in the process” from nursing unit to nursing unit and, in turn, deduce from the report data “the impact those variations have.” All in all, Paoletti notes that AU MEDS presented Lancaster General with “a great opportunity to share best practices from one nursing unit to another. That is a tangible benefit.”

Trumping the Skeptics
“At first,” Ferdinand admits, “I thought our nurses would do things correctly when they had an observer, instead of performing their actual process. But interestingly enough, that’s not true. Because the nurses don’t necessarily realize they’re making a mistake, they go through their normal processes.” In addition to Lancaster General’s non-punitive reporting environment, Ferdinand notes that because AU MEDS was introduced along with the Bridge system, “our nurses see it as a way to improve our bar coding process, not as a way for us to hunt for their errors. They know we want to identify and subsequently minimize the risks in the process.”

Ferdinand looks forward to a time when the use of AU MEDS is more widespread, as users can then share their data, results, and best practices. Of course, there is something intimidating about a system that details just how high your medication-error rate is. According to Ferdinand, “the data hits you in the face; there are huge opportunities for improvement. It is just up to the institution to use the information to start improving.” Suess adds, “People need to be willing to try it and see the benefits. There is a wealth of data out there that we can learn from.”

Before and after their Bridge implementation, Lancaster General used AU MEDS to assess their processes and practices and the impact bar code technology would have on patient safety. The introduction of bar code technology led to a 54% reduction in medication errors. The use of AU MEDS’ direct observation provided Lancaster General with a clear list of processes that would be corrected with the introduction of bar code technology, and spotlighted processes that bar code technology would not impact. This provided a clear indication as to which practices and processes needed to be changed for both pharmacy and nursing.

Future Plans for AU MEDS
Because of the success they have had with AU MEDS system on the nursing floor, Paoletti hopes Lancaster General can “use AU MEDS to observe our pharmacy dispensing process and ultimately improve our efficiencies and medication turnaround in the pharmacy.” Suess thinks there will also be a role for AU MEDS in Lancaster General's specimen-collection procedures, “because we know there are a number of specimen errors that we are not catching.”

A Collaborative Approach
To maximize the system’s benefits, the implementation of AU MEDS should involve both pharmacy and nursing. Ferdinand contends, “If you keep pharmacy out of the loop, you can only impact the errors that are related primarily to nursing. Some of the issues the system will identify are clearly owned by nursing, and some are clearly owned by pharmacy. However, because most of the issues impact both departments, they have to work together.” Further stressing the importance of collaboration between pharmacy and nursing, Paoletti adds, “They not only have to work together, but both also have to be committed to expeditiously resolving problems and to improving the medication-use process.”