More and more hospitals looking to end drip-medication delivery errors are clearing their existing fleets of infusion pumps and replacing them with so-called smart pumps, the latest in technology.

Smart pumps work much the same as their nonintelligent predecessors, except that everything is managed by an onboard computer system. The computer runs programs to ensure the right drugs are delivered to the right patient, in the right amounts, at the right time.

This concept that four rights do not make a wrong is good news where patient safety and satisfaction are concerned. However, translating the promise of smart pumps into reality requires sound strategic planning and execution.

“Smart pumps are one of the most complex and challenging implementations that most institutions will ever go through,” warns W. Glenn Scales, CBET, a patient safety specialist in the department of clinical engineering for the Duke University Health System, Durham, NC, which a few years ago deployed a total of 4,500 smart-pump bases and software modules.

What makes deployment so daunting is, first, the need to coordinate the rollout across many different departments—physicians (who are represented from all sides such as oncologists, surgeons, etc, and have different needs), nursing, anesthesia, the operating rooms, ICUs, and general floors, as well as the pharmacy, information technology, and various administrative units (which may or may not include the hospital’s attorneys because of the possible medicolegal implications of this particular device).

According to Scales, this is new territory for biomeds because working with a multidisciplinary group becomes an enterprise project as opposed to an installation they can handle on their own. In addition, the devices need to interface with the electronic medical record. Another challenge that needs to be addressed is keeping up with security protocols.

Moreover, there can be no deployment until such time as a quantity of smart pumps have been purchased, which means shopping expeditions, price haggling, and the lining up of various critical support services.

Someone, of course, must be put in charge of all this. Often by default, the burden falls to a leader from nursing or the patient safety department. But there is no reason why the ball could not just as easily (perhaps easily is the wrong word choice) be carried part or all of the way by a biomedical/clinical engineering department.

“Although it may appear that the role for clinical engineering doesn’t begin until the devices arrive, nothing could be further from the truth,” Scales says. “There is a huge benefit to engaging clinical engineering at the very beginning [of the planning process]. The reason is that many of the decision-makers will be physicians, nurses, and pharmacists who understand the clinical applications of the device but not the technology behind it. So it is helpful to have someone with technical expertise explain to them the hardware and software functions and the device capabilities [in order that they may make better-informed decisions at key junctures].”

No disagreement on that from Glenn R. Perry, CBET, manager of clinical biomedical engineering at The Queen’s Medical Center in Honolulu, home now to some 800 Hospira Symbiq smart pumps deployed earlier this year. “Biomed needs to wedge itself into the beginning of the process,” he says.

By that Perry means biomeds should strive to establish their value as a resource for a smart-pump conversion. Start, he suggests, by reviewing and monitoring the service records of existing pumps and, “using your metrics,” bring this data to nursing, safety, and purchasing to get the interest of the directors of those areas. As always, talk with the users and the managers on the clinical floors to find out what problems they are having. This will give you an idea of what needs the users will be asking for in terms of smart-pump capabilities. “Then, ask to participate in safety committee meetings to see where the clinical users’ focus and objectives are,” Perry says. “Be prepared to describe—in terms of patient safety—where you see improvement needed. Bring up details of warranty, service cost, uptime, and operation that you see getting in the way of success.”

According to Scales, the important focus for biomeds is to work on building a strategic plan within a multidisciplinary group.

A Multidisciplinary Approach

Hospitals typically begin talking about smart-pump conversion after frets over medication-delivery errors reach a certain high level. At Rady Children’s Hospital in San Diego, Chris Abe, RN, BSN, CIC, HEM, senior director of safety and support services, says such concerns prompted her institution in 2002 to ramp up for a smart-pump conversion. “Infusion smart pumps were a relatively new innovation at the time, but I knew they could be a solution worth exploring,” she says.

A multidisciplinary task force undertook the exploration. “Every clinical area in the hospital was asked to appoint a representative to serve on our implementation group,” Abe says, who also heads up the biomedical engineering department. “We ended up with an implementation group of 26 members, consisting of not only nurses but also pharmacists and physician champions.”

The group was structured in teams of threes (a nurse, a pharmacist, and a physician), which were referred to as triads—one for the NICU, one for the PICU, another for the step-down unit, the med-surg section, the operating rooms, and all the rest. Each triad and its component members were assigned well-defined roles. “The triads would develop the safety components for their respective clinical areas, then during regularly scheduled full group meetings, share with members of the other triads what had been accomplished,” she says. This way, all teams—even as they toiled apart from the others—stayed plugged in to the overall process and were able to exchange ideas or solve challenges in common.

Smart Implementation Strategies

When it came to the actual implementation of infusion smart pumps, things went well at Oregon Health Sciences University. Dennis Minsent, MSBE, CCE, CBET, director of clinical technology services there, credits the success to the many different players who were involved—including the vendor for supplying “an entire cadre of people” to help change tubing from the old pumps to the new and to assist with the training of the hospital’s staff.

Currently at Minsent’s institution, smart pumps are deployed “everyplace where you would expect to find an infusion device in operation,” he says. “We have 615 Alaris main programming modules and 1,410 general-infusion pump modules.”

The hospital’s critical care tower will soon expand by 60 beds; accordingly, more smart pumps will be rolled out. “We’re looking to acquire additional modules to go with the base units,” Minsent reveals. “In particular, we’re giving consideration to the PCA, end-tidal CO2, syringe pump, and bar coding modules.”

Duke University Health System, Durham, NC, likewise reported a generally smooth rollout. Of the few minor bumps along the way, the one that prompted the deepest furrowing of the brows centered around the screen view of the drug library data sets. W. Glenn Scales, CBET, patient safety specialist in the department of clinical engineering, explains by way of a cautionary note: “The formatted appearance of the data sets on your library development computer screen as you develop and review data using the supplied management software is not necessarily going to be the way they appear on the pump’s display screen. Yet, how the data are presented on the pump screen is where the rubber meets the road. In our initial rollout of smart pumps, we did not realize there would be any difference in appearance. Later, when [expressions of displeasure] started coming in from the clinical users, we had to go back out and tweak the formatting.”

Implementation at Rady Children’s Hospital, San Diego, saw that institution mothball 300 traditional pumps and in their place bring aboard 1,150 smart pumps. “During the initial month after rollout, as nurses were becoming accustomed to the pumps, there were quite a few user errors [pegged to unfamiliarity with the equipment],” says Jay Damasco, a level III biomed. “We were very early adopters of the smart pump, so we also encountered problems with flawed component design on the boards. But those were addressed, and since then we’ve had mostly trouble-free operation.”

Today, he says, “Everyone here appreciates the modularity of the pumps. Previously, with our traditional pumps, we had just three channels and none could be moved around. So, if one channel didn’t work, the entire pump had to be taken down. Now, with these smart pumps, if something fails, you just take off the affected module and pop in a new one, without losing the ability to continue using the pump itself.”


Perry cautions to expect impatience at just about every step of the way toward smart-pump deployment. “Everyone wants the implementation to be a success, but their tendency will be to hurry through it,” he says, adding that the biomedical/clinical engineering department can help matters by riding herd over this tendency to speed through.

Sometimes, though, speed is of the essence, as was the case for Oregon Health Sciences University in Portland. “We made all the key decisions regarding smart-pump implementation planning within a 5-month period during 2005,” says Dennis Minsent, MSBE, CCE, CBET, director of clinical technology services. “We were under pressure to complete the process as soon as possible so that our deployment would coincide with the opening of a new critical care tower here on campus.”

An initial step involved establishing goals for the use of the smart pumps, Minsent recounts. Improved patient safety was the obvious goal. “But, beyond that, we also wanted to be able to utilize quality assurance data from the pumps to develop greater insights into bedside practice.”

Another important goal is to procure vendor support. Perry believes such support should be broad and deep, and should involve vendor sales, service, and applications representatives, beginning months ahead of the planned deployment. Says Perry, “Emphasis should be placed on having the vendor provide training and periodic refresher training to the users.”

Network Issues

Among the vital components of a smart pump is its self-contained library of intravenous drugs. This library guides the dosing of medications at bedside and warns when a dosing error appears imminent or is in progress. This library is one that can be (and usually is) developed by the institution’s own pharmacists and physicians working in cooperation with one another. Before this happens, though, clinical/biomedical engineering—in its leadership role—should acquire a thorough understanding of the mechanisms or processes by which the completed drug library will be loaded aboard the pumps and then periodically updated, Perry recommends. “If possible, see how this is done at other hospitals that are using the same technology, and then be able to demonstrate the full process in a trial area of your own hospital,” he says.

Networking is another issue that ought to be addressed during planning. “You’re going to be bringing in devices that likely will connect to the hospital’s IT network,” Minsent says. “Consequently, data security and HIPAA are key elements that need to be considered, especially if you plan to use wireless protocols.”

Minsent, therefore, finds it imperative that biomeds consult the expertise of the IT department at the earliest imaginable stage. “IT needs to be part of the technology identification, evaluation, and selection process,” he says. “The last thing you’ll want is to press forward without IT involvement and bring in a fleet of smart pumps only to be told later by IT that, because of inadequacies inherent to the pumps’ security software, or because of the need for unusual configuring, you won’t be allowed to operate the devices on the hospital network.” This would represent a blow that would negate much of the smart-pump advantage.

Adds Minsent, “There are many organizations that struggle with their IT departments in getting authorization to run the device on a wireless network.” In some instances, IT opposition arises from concern over bandwidth consumption, although Minsent believes that objections based on that are in large part groundless. “Typically, with these pumps, they don’t send and receive data continually,” Minsent explains. “They transmit only when polled, and they receive only when you push data out to them.” In other words, smart pumps are unlikely to slow net traffic by much.

Nonetheless, Perry thinks it is a good idea to work with IT on developing a phase-in plan for introducing those pumps to the network. “This will help find any bottlenecks and issues with signal strength, bandwidth, and all the rest,” he contends. Similarly, plans should be made to trial a small number of smart pumps with a select group of users before rolling it out across the enterprise. Glitches, Perry says, can be more easily caught and rectified this way—and the lessons contained therein can be applied to good effect, enterprisewide.

Rich Smith is a contributing writer for 24×7. For more information, contact .