Published in the April 2011 issue of Today’s Hospitalist
WHEN IT COMES TO BAR-CODING INITIATIVES to reduce medication errors, the medical literature and mainstream press are full of stories about how the technology can backfire. Bar-coding scanning has the potential to significantly reduce medication errors, but hospital staff sometimes use dangerous workarounds to circumvent the technology. In worst-case scenarios, errors can actually increase.
Baystate Medical Center in Springfield, Mass., was more than aware of those potential downsides. But rather than shy away from technology that it knew could help, the hospital embraced bar-coding head-on. In the process, it slashed its medication error rate by 75%, according to its own self-reporting event system.
That’s not to say that Baystate didn’t encounter its share of challenges. The number of non-serious medication errors actually spiked briefly, especially of "near-miss" events. But in the end, Baystate succeeded in rolling out the barcoding initiative to more than 20 nursing units in just a matter of months.
Baystate tackled bar-coding as part of a several-year initiative to improve medication safety by leveraging technology. The hospital, for instance, had recently implemented a new clinical information system that included advanced CPOE.
The center first piloted the use of bar codes in April 2008 on both surgical and medical nursing units. A nurse uses a hand-held computer to read the bar code on both the patient’s wristband and the medication label. An alert is displayed if it’s not the right drug in the right dose via the right route to the right patient at the right time.
Early in the process, the hospital encountered problems scanning some medications. Thomas Higgins, MD, MBA, interim chair of the department of medicine and the initiative’s physician lead, says that medications “particularly those wrapped in foil, or in packages that can crinkle or smudge “could be hard to scan. Scanning large volume parenterals also presented challenges.
In addition, alerts would go off if a physician order did not exactly match in terms of route of administration or product form, such as a capsule vs. a tablet. That led to alert fatigue and frustration.
"Because there must be an exact match between the ordered and delivered medication, some alerts fired even when care was correctly delivered," Dr. Higgins explains. "With multidose insulin vials, a bar code won’t necessarily reflect what you’ve drawn up." Similarly, if vancomycin had to be delivered enterally for C-difficile colitis, the system would balk.
Certain frequently administered products, including IV saline solutions and suppositories, also confounded the system. Baystate decided to purchase its own bar-coding equipment to create codes for items that didn’t already have one. Products that the hospital buys in bulk, for example, often aren’t bar-coded.
Not all of the problems were the result of system glitches. Beverly Siano, RN, the director of adult hospital care and operations who co-led the initiative, recalls another unanticipated challenge: scanning bar codes in patient rooms at night, when the lights are dimmed or turned off.
Because patient bracelets had only one bar code, nurses would have to maneuver the bracelet to scan it. As a result, they were disturbing patients to administer an IV drug or replace a saline bag. To make scanning easier, Baystate later added multiple bar codes to the bracelets.
Gary Kerr, PharmD, MBA, director of pharmacy services, says that the hospital was also very proactive in addressing workflow issues being encountered by the nurses.
"So much of this is process and workflow at the bedside," Dr. Kerr says. Scanning rates drop when products aren’t scanning correctly, for instance, or if “as some organizations have learned “the whole process is so cumbersome that nurses are tempted to devise dangerous shortcuts, like scanning bar-coded labels instead of patient bracelets. Project leaders made sure to educate staff about the danger of workarounds, and they instituted a zero tolerance policy for such practices. According to Dr. Kerr, pharmacist-nurse observation teams also made sure that scanning was being done correctly.
Between spring 2008 and early 2010, the center’s self-reporting event system reflected a drop in medication errors from 1.2 per 1,000 patient days to 0.3 per 1,000 days. That represented a 75% reduction in errors reaching the patient. (Experts estimate that about 2.5% of medication errors are dispensing errors, but that number may be low because so many errors go unreported.)
But that success came after the hospital addressed another problem. At baseline, self-reported dispensing and administration errors were 20 per million doses dispensed. However, the bar-coding system was picking up many dispensing errors that previously would never have been reported, such as finding the wrong or outdated medications in cabinets. After the initiative began, that rate rose to 38 per million doses dispensed. By 2009, however, only five errors reached the patients, a rate of just 0.67 per million doses.
The bar-coding initiative was so successful that Baystate introduced it in critical care units three months later and rolled it out to all 20 nursing units just three months after that. The initiative has since been expanded to the several facilities in the Baystate health system. Baystate’s investment in initiative software and hardware came in at just over $385,000.
The "big bang" approach
According to Ms. Siano, one key to success was encouraging nurses to talk about near misses. "Our nurse champions were on a mission to find those stories and report them, so people understood first-hand the value of barcoding in improving patient safety," she explains.
And while everyone has heard about bar-coding initiatives gone wrong, Ms. Siano says that in many of those cases, end users across departments weren’t engaged early on in planning and design.
"When you talk to other organizations that have problems, you hear about bar-coding being driven by only one department," she says. "Planning and leading such change must be a collaborative process."
Dr. Kerr admits that the initiative might have initially gone more smoothly if the hospital had set less ambitious targets. Baystate’s goal was to have 100% of patients scanned (and 90% of medications scanned) within three months. "I think we’d set those goals lower if we did this again," Dr. Kerr explains, so that staff could see progress more readily.
Overall, however, project leaders were happy that they went with a "big bang" approach, rather than the sequential rollout that was initially planned.
"The staff decided that we were ready," Ms. Siano says. "Bar-coding closed the loop and created a safety net within the medication management process."
Bonnie Darves is a freelance health care writer based in Seattle.