Hospital pharmacy helps design specialized IV medication delivery system

Hospital wins national award for system

Hospital pharmacies sometimes accept new technology without having a say over its design and ability to fit into the hospital's workflow. Instead, staff will develop work-arounds and find cumbersome ways to make the new systems work.

There's another option to this dilemma. Instead of making your work system fit the new technology, you can make the technology fit your work system.

Pharmacists at Kaleida Health, Women & Children's Hospital of Buffalo in Buffalo, NY, did exactly that when they realized the new IV pump technology was not working out as well as planned.

"When we got deep in the process we realized all the dosing capabilities for that pump for the health system wouldn't work for our pediatric population," says Kelly A. Michienzi, PharmD, clinical pharmacy coordinator for the hospital.

The hospital implemented a process improvement program that resulted in a better and safer IV medication delivery system and earned the hospital a national award. The American Society of Health-System Pharmacists (ASHP) gave the hospital an Education Foundation Award for Excellence in Medication-Use Safety in December 2009.

At first, pharmacists noted the work-arounds suggested by the vendor, making adjustments as needed. But as these increased, their comfort level decreased, she notes.

"It was too alarming," Michienzi says.

"When we compiled all of our concerns and presented these to the pharmaceutical and therapeutics (P&T) committee, then the hospital administration went to the health system administration, and everyone agreed we'd have to put the project on hold at this site," she explains.

"The administration backed us and allowed us to explore other options," Michienzi says. "The new pump we selected was not yet marketed, so we had a lot of input on its appearance and design."

Michienzi, a physician, a nurse educator, and the hospital's quality director compiled information to present to the P&T committee. They listed specific drugs, including some high-risk products that could not be dosed in the conventional way through the new pump because of software limitations.

"We showed people the work-arounds suggested to us and showed them how confusing it was to staff and how it could lead to errors," Michienzi says.

There were 10 or more drugs on a spreadsheet that showed how incorrect dosing could result in adverse outcomes.

The drugs were dosed on a per-kilo basis. The pump had a maximum adult dose listed in the system, and this maximum would appear even when the pump was being used for children, so someone would have to manually adjust it for a pediatric population, she explains.

"At that time, the vendor made you select either adult or pediatrics data, and we needed both," she adds.

Once the project was put on hold, a team of biomedical engineers, nurses, nurse educators, physicians, information systems staff, the director of pharmacy, and Michienzi met to look at products offered by other vendors and other product lines by the same vendor.

Since they'd learned from the first abortive experience with new pump technology, they asked more direct questions when reviewing other products.

After several meetings, they picked a product, reviewed its software capabilities in terms of dosing conventions, and assessed its compatibility with the hospital's network capabilities for wireless connection.

"Continuity of flow was a big issue," Michienzi says. "We needed a small infusion rate, and we wanted to ensure the continuity of flow was good at small infusion rates."

A first step was to start building the system's drug library.

"The vendor would bring revisions to us for feedback, and we'd have an opportunity to interact a lot with the vendor to make features that were desirable to us," Michienzi says.

The drug library was created based on the literature and practice standards.

"We'd take it out to the nursing staff, nurse educators, and staff with a lot of experience to ask for feedback," she says.

"We'd ask, 'How fast do you usually get this drug? Does it happen in over 2 minutes?'" she recalls. "We'd make sure the limits matched actual practice and the guidelines, as well."

Initially, there'd be a remote drug library server, and Michienzi could log onto the vendor's remote site and input data.

"Once we owned the software we were able to transfer data into our own server, and eventually we had separate pumps and could load this on the pumps for nurses to practice on," she says. "We developed a relationship with the information systems people and biomedical engineering people when previously we had not had a lot of opportunities to work with those groups."

Installing the IV medication delivery system effectively required a great deal of trial and error, she notes.

"We'd ask nurses to pretend they were pumping this for a patient to see what they were supposed to do and make sure the pump would let them do it," Michienzi says.

These informal pilot sessions produced feedback that Michienzi used to make revisions in the drug library. She found that nurses were used to being creative with technology, finding ways around perceived obstacles that she might not have imagined.

"Then I'd have to go back to the drug field and put in limits in other fields to prevent the work-arounds," she says. "We must have repeated that process at least 20 times."

Finally, the limits matched the practice and references, and the work-arounds ceased.

"If there were any times when nursing and pharmacy didn't agree on a drug dose limit then our P&T co-chair who is a toxicologist would say, 'This is the limit where you need to stop because it would hurt someone,' and he'd put in hard limits that couldn't be overridden," Michienzi says.

The revised system was pilot-tested in the pediatric and neonatal intensive care units and the anesthesia areas first because most of the IV medication delivery drugs were used in those areas.

"These were the super users who received additional training on the pump," she says. "During those training sessions we had more nurses working on it for longer periods of time."

These super users continued to find little problems that were tweaked in the system.

The system now has pumps communicating information wirelessly into the server, and pharmacists and other staff have a variety of report options available. They can see which pumps have had the most recent version of the drug library, where these are located, and they can see whether employees are using the drug library.

"We can check dosages on the pump, how often there are overrides and what percent variance," Michienzi says. "And we can par that information down by specific units or to the whole hospital."

They use some report information to change the library's drug limits. For instance, if a drug limit is too low or if it's too high and no one is hitting this limit, then they can use this information to check drug library compliance, she adds.

"We've actually changed some dosing limits," she says. "For example, we made two different entries for naloxone, using lower limits for patients who are using it for pruritus or itching from narcotics and higher limits for people using it for narcotics reversal if an overdose was intentional or otherwise."

They discovered from the pump's reports and incidence reports that if someone misprogrammed the rate, then there was a possibility that someone who only needed the drug for itching could receive the higher dose for a narcotic overdose.

"So we made two separate entries in the drug entry, one for the higher dose and one for lower dosing," Michienzi says.

Reports also made it clear which type of errors the pumps prevented. For example, a nurse who hit a limit might go back and change the dose in the pump, and this would result in an error report.

"If a nurse programmed it for too much then we took that as a missed error," she says.

Overall, the hospital staff was satisfied with the new IV medication delivery system, giving good feedback on the final result, she notes.

"I've had one PICU nurse manager tell me that if we tried to take the pump away from nurses we'd have a fight on our hands," Michienzi says.