The use of smart infusion pumps improves the quality and safety of patient care, given their ability to deliver a specific infusion rate for narrow therapeutic index infusions to simplify rate control and to promote accurate medication delivery at the bedside. The inclusion of dose error reduction systems (DERS) offers the capacity to program customized infusion limits into smart pumps for each medication and customize limits for the same medication by care area. With this technology, pump programming errors due to incorrect data entry or a lack of knowledge are minimized; nevertheless, the full safety benefits will only be realized if the changeover process is well planned and thoughtfully executed. 

Smart Pump Considerations
Mercy Medical Center in Sioux City, Iowa implemented smart pump technology in October 2011 as a result of an FDA-mandated recall of the traditional infusion pumps then utilized at our facility. Organization-wide collaboration was integral to the ultimate purchase decision and implementation. Major considerations included pump type, infusions required in different patient care areas, dosing limits for various care areas, available concentrations, infusion naming conventions, and staff training. 

Several types of smart pumps were available in a variety of configurations, including pumps with software and drug libraries updated wirelessly or via a wired connection, single or multiple channels, and audit trail availability. Deciding whether to purchase a wired versus a wireless system is a crucial consideration. It is important to determine whether the facility can afford the resources necessary to locate each IV pump, plug it into a computer, and upload a new update to the drug library every time a drug product change is required. If this is the case, then a wired pump may be an appropriate choice. For larger pump fleets, the wireless option negates the necessity to physically find each pump and connect it to a PC to conduct the update. 

The utility of single versus multichannel pumps is generally a factor of the patient population served. Single channel devices are the most efficient choice for patients receiving only one infusion. Multichannel devices best serve areas such as critical care settings where multiple infusions are common. Additional benefits of utilizing multichannel devices are the reduced QA time for biomedical staff and the need for fewer IV poles. 



Implementing Smart Pump Technology
Pharmacy became the project champion and lead the implementation process with the input of nursing, administration, biomedical services, central supply, and the IT department. The first step was to determine the quantity of pumps needed. In order to gather pump usage data, a line assessment team was assembled to check every pump throughout the hospital in a single day, noting how many infusions were occurring via pump and how many channels were in use on each pump. These numbers were compared with the patient census and a calculation was made to best determine how many pumps would be needed if the census was at its maximum. The team’s final determination indicated that 237 single-channel pumps were required. 

The primary goal in implementing the new pumps was to maximize patient safety. As such, nursing and pharmacy collaborated to decide how to configure and program the pumps and create the drug library. Determinations including care area nomenclature, IV fluid specificity, default keep vein open (KVO) rates, alarm volumes, and availability of code blue infusions in non-critical care areas were made, in addition to an overall policy review to determine which infusions would be most appropriate for differing care areas.

Establishing Protocols
While the pump programming process was straightforward, determining the appropriate protocols and limits was challenging. When an infusion is initiated, the bedside clinician is prompted to choose a care area. Each care area represents a patient population or patient care unit and the drugs listed in that care area can have limits customized to that area. Ideally, care areas should contain only the medications needed for a particular patient population or unit (eg, medical-surgical or ICU). For example, a telemetry floor or ICU pump may be programmed to infuse 20-40 mEq of potassium chloride in one hour, while the medical-surgical floor would not be permitted to infuse potassium at a rate greater than 10-20 mEq per hour in an effort to minimize venous irritation or avoid negative cardiac effects.

Identifying and naming the care areas was required before the pumps could be properly programmed. One approach is to create an individual care area for each nursing floor. Thus, the care area could be selected based on the physical location of the patient; the disadvantage is that whenever a patient is transferred, the pumps require reprogramming to reflect the patient’s new location, increasing the likelihood of programming error and costing the bedside clinician additional time. Thus, the alternative was adopted, wherein a limited number of care areas were assigned based on larger groups of patient populations. For example, the advanced care area encompasses the ICU, CCU, ED, and non-OR procedure areas. If a patient is brought from the ED to the ICU, the nurse does not need to reprogram the pump to ensure the limits are congruent with the level of care available.

Effective care area construction can streamline pump programming, but keep in mind that it requires multidisciplinary input. In order to determine the appropriate drug products for each care area, the team reviewed facility and unit infusion policies with nursing administration. A properly constructed care area will include only medications that may be utilized in that area, thus reducing the potential for programming error due to the selection of an incorrect product while promoting overall infusion safety. Mercy Medical Center created a total of six care areas: advanced care, anesthesia, medical/surgical/OB/rehabilitation, oncology, pediatrics, and six post (a step-down unit).

In determining IV fluid nomenclature, some members felt that all IV maintenance fluids should be represented individually, but this would create a large drug library that would make selection of specific IV fluids tedious for nursing staff. Of particular concern was the inability to include all possible combinations of nonstandard IV fluids. Therefore, a broad definition of IV fluid, utilizing two IV fluid selections, was selected: IV fluid with potassium or IV fluid without potassium. This definition requires less pump programming, simplifies the drug library build process, accommodates all types of IV fluid products, and provides infusion safety based on potassium content.

Other parameters were designed to maintain consistency with facility standards. For example: 
n Infusion product names were matched with the pharmacy system and the eMAR, ensuring the product name in the library matches the pharmacy label 

• The KVO rate was set to align with pre-existing nursing policy 
• Alarm volumes were set and locked to the high setting for all pumps in all care areas. This setting was chosen to prevent staff from adjusting alarm volumes, causing the potential for an alarm to go undetected 
• Secondary call back alarms were configured so that bedside clinicians would be prompted to activate or not activate these alarms for use with IV piggyback infusions. This allows bedside clinicians to retain autonomy with these alarms based on their own nursing practice

Drug Library Development
Building the drug library was the most time-consuming and clinically challenging part of the smart pump implementation process. The build was based on clinical knowledge, available evidence, and existing infusion policy and practice. Colleagues at other organizations shared information from their IV pump drug libraries, providing us with preliminary infusion data, and we developed our own default infusion rates and dosing limits based on drug libraries of other facilities, local nursing practice, and tertiary references. These comparisons were completed to ensure accurate and well-accepted infusion practices and minimize unnecessary alerts. We designed default infusion rates and dosing regimens to be in agreement with default pharmacy system parameters and standard concentrations. This eliminated the need for bedside clinicians to convert flow rates from one set of units into another. All of these considerations promote infusion consistency, reduce programming time, and subsequently minimize the opportunity for keying errors. 

Soft and hard limits, both high and low, as well as default rates, were developed by pharmacy, anesthesia, nursing, and other affected clinicians based on an aggregate of data. Hard limits are those that the clinician cannot override; soft limits trigger a warning, but may be overridden by the bedside clinician. Usual infusion limits were abstracted from tertiary references, policy and nursing practice, drug libraries from other facilities, and clinical judgment. Limits are customizable by patient care area. For example, limits for vasopressor drips, propofol, and antiarrhythmics are generally more conservative in the ICU and ED than in the OR; this is due to the intensity of infusions utilized in operative settings and the fact that nurse anesthesia staff and other providers are directly managing infusion therapy. The pharmacy and therapeutics and medical executive committees reviewed and approved the soft and hard limits and default infusion rates.

Chemotherapeutic agents present a unique set of challenges given the changing research protocols, thereby making standardized volumes, infusion times, and concentrations continuously moving targets. These medications often create the greatest risk for potential harm, but they cannot be standardized to the same degree as other infusions. To build the chemotherapy drug library, we reviewed approximately one year of chemotherapy infusion data, and using database software, constructed drug files based on a typical range of infusion volumes and rates. For example, the clinician can choose from several etoposide files. One was built to be infused over one hour, and others over two hours, four hours, or over 24 hours. Volumes infused over the previous year were incorporated into the programmable volume and rate range in order to create acceptable limits. If a chemotherapy volume is too high or too low to be infused over a specified time, a soft or hard limit will be triggered notifying the clinician of a potential issue. Although this system is not as standardized as non-chemotherapy infusions, it provides a distinct advantage over completely circumventing DERS.

Access to Emergency Infusions
To ensure emergency infusions are available to every patient in the facility, all care areas include a separate roster of critical care infusions. These infusions are listed as “ZZ“ drugs (eg, ZZ-epinephrine). If a code occurs on a general medical unit, a bedside clinician can search “ZZ” and choose from a list of vasopressors, antiarrhythmics, rate control agents, and related products without having to change care areas in the pump library. The goal of the “ZZ” drug list is to ensure rapid availability of emergency agents without requiring guesswork on the part of a bedside clinician, who is already functioning outside of a procedural/critical care setting. 

If a drug is not found in the drug library, the basic infusion mode remains an option as the clinician can operate the pump while circumventing the DERS. However, this practice is discouraged, as it negates the safety benefits the DERS provides. Bedside clinicians are educated that using basic infusion mode should occur infrequently, and in instances when it is necessary, they are to call pharmacy for guidance or to request a drug library update. Requesting an update is straightforward: nursing staff simply completes a pharmacy-developed IV pump change request form, which is available on the facility intranet page, and forwards it to pharmacy. Clinical pharmacy staff reviews all requests and follows up with clinicians. If drug library changes are required, at least two pharmacists review the change before the update is completed. These updates occur on a quarterly basis, or more frequently if an emergent change is required.

Rolling Out the New Pumps
After all the preparation steps were completed, the physical pump changeover began. Two multidisciplinary teams, including super-user nursing staff, pharmacy, biomedical, transportation, IT, and vendor support, were organized to complete the changeover process. The vendor provided support staff during the week of implementation to conduct training for all staff and afford 24/7 on-site and on-call bedside clinician programming support. The bedside clinician training sessions delivered basic information on loading the IV sets, searching for drug products, navigating pump options, and general pump operation. 

The evening before implementation day, the old pumps were removed from surgery and procedural areas and replaced with the new smart pumps. The following morning, implementation began in the ED, so that any patient admitted from the ED would come to the floor with a new smart pump. The two teams then began at the top and bottom floors of the facility and went room-to-room changing over pumps until all pumps in the facility were replaced. The changeover process occurred over approximately four hours.

The entire smart pump purchase and implementation process took nine to 12 months to complete. Thanks to excellent multidisciplinary collaboration, the drug library required only a small number of programming adjustments after implementation. On implementation day, an issue was encountered with the pumps and the local wireless network that was interfering with drug library updates. Thanks to the collaboration between the hospital’s IT staff and the vendor’s IT support, the issue was identified and resolved in under an hour. As the implementation progressed further, it became apparent that radiology technicians and staff frequently needed to silence pumps or temporarily suspend infusions for radiology procedures. Pharmacy added go-live training for the radiology staff so they could competently operate the pumps during radiology procedures.

Conclusion
Smart pumps are a valuable patient safety advancement. The data available from reporting software has allowed our facility to fine-tune overall pump utilization and drug library design, while identifying opportunities for clinician education. We look forward to future enhancements, such as bidirectional wireless communication, to provide real time documentation of infusion data and other advanced pump technology to further enhance accurate medication delivery and safety at the bedside. 

The American Society of Health-System Pharmacists publishes an informative book by Pamela Phelps entitled, Smart Infusion Pumps – Implementation, Management and Drug Libraries, which details the entire process of implementing smart pump technology. We found this reference particularly helpful. 

Brett Bieber, PharmD, BCPS, is a member of the clinical pharmacy service and pharmacy systems team at Mercy Medical Center in Sioux City, Iowa, and serves as adjunct faculty in the pharmacy technician training program at Western Iowa Tech Community College. He is a board certified pharmacotherapist and received his PharmD from Creighton University. Brett’s professional interests include adult medicine, antimicrobial optimization, medication reconciliation, and the education of the next generation of pharmacy students, residents, and technicians.