Safety First: Smart Pumps in Action

March 2009 : Advancing Patient Safety - Vol. 6 No. 3

Click here to visit the ProCE website for this program

To improve safety in the medication administration process, addressing medications that are delivered intravenously (IV) and which provide the highest risk to patients can result in life saving critical catches. At our hospital, the Medical Center of Columbus Regional Healthcare Center located in Columbus, Georgia, we sought to address and correct these IV medication management issues through the implementation of intelligent infusion devices. Our goal in using intelligent infusion devices, or smart pumps, was to improve the quality of care by establishing safety limits and standardizing optimum administration processes.

The Medical Center, a 413-bed community teaching hospital, offers highly specialized acute care services, plus the region’s only advanced maternity services and a neonatal intensive care unit, a full service pediatrics program, and a pediatric intensive care unit. We also are a Level II trauma center, serving 13 Georgia counties and offer many other medical-surgical services, including a large oncology center. We have a large family practice residency program, and this year, we had nine pharmacy residents. Additionally, we are a disproportionate share hospital operating in one of the lowest income areas in the state of Georgia; therefore we run an extremely thin net margin.

Pump Features
In 2006, we implemented intelligent infusion devices in our institution. These infusion devices provide precise delivery of medication, and when configured with their software system, offer a drug library containing dosing limits with the potential to prevent severe or even fatal medication errors at the bedside. With the correct use of smart pumps, we can now ensure greater safety in the IV route of drug administration.

Some basic administration features we looked for in our smart pumps were positive valving, set based free-flow protection, an air-trap management system, and back priming for air elimination. Additional features we value are automated piggyback bag delivery, automated concurrent delivery, keypad lockout via toggle switch, programmable standby settings, multi-step delivery, loading dose automation, and programmable delayed starts. The pump can be programmed to use exact dose, rate, and volume parameters for specific medications. TALLman lettering specific to each institution is employed to distinguish between look-alike and sound-alike drugs. On the final display screen, the clinician is asked to confirm medication type, dosing, and variable rate information instructions for entries on both lines A and B, so that primary and secondary infusions are under the umbrella of the medication safety limits. User-defined soft and hard lower and upper limits for each medication are programmed into the drug library. If the medication is being dosed without limits in operation, a special symbol appears on the confirmation screen visible to the clinician before final confirmation is requested.

Drug Library
One of the most important steps in the successful use of infusion devices is the creation of a comprehensive drug library. The library must contain the safety data for all IV medications used, including name, concentration, minimum and maximum infusion rates, and minimum and maximum bolus infusion rates per specified time, all with appropriate dosing. The specifics of dosing and infusion rate can be customized to the hospital’s individual clinical care areas (CCAs).

Initially, in our hospital, we started with broad CCAs, such as a med/surg grouping that included multiple patient care areas. We soon realized these groupings were too broad as they did not allow us to track an individual unit’s compliance rates or use of the pump. For example, usage varied greatly between the emergency department and a standard med/surg unit, but we were not able to capture these differences as they were grouped in the same CCA. As a result, we chose to customize the libraries to individual patient care areas, such as pediatrics, neonatology, outpatient cancer center, and specific ICUs. The pump provides storage of individual drug libraries for up to 18 CCAs, allowing us to track an individual unit’s pump usage, which greatly increases the value of the data we collect. Each library contains user-defined dosage limits, both soft and hard, based on clinical evidence, literature, and institutional best practice standards that can be updated and accessed via the software.

Establishing the safety limits within the drug library to meet the patient needs within each CCA is critical in order to maximize patient safety. I recommend establishing both soft and hard upper and lower limits. The soft limits drive system alerts that act as warnings, which the user can override or edit with confirmation. The hard lower and upper limits cannot be overridden. There is a “no drug selected” menu option on the pump that allows for infusion without programmed limits. This may be used, for example, in the case of an emergency administration of critical drugs initially infused at a very high rate, such as norepinephrine. These hard and soft limits, both upper and lower, are at the heart of the pump’s ability to avert fatalities and must be tailored to meet the needs of each individual CCA in your institution.

The software system was equipped with a drug library template that we proceeded to customize. We consulted the literature and asked other institutions to share their drug libraries with us. We discovered that many hospitals are not assigning safety limits in their library, often due to internal disagreements over what the number should be. Our own experience has taught us that establishing some limit is better than no limit in terms of quality patient care. At the very least, a hard upper limit — even one that is set very high — can preclude a triple digit entry when a double digit was intended. While the different pumps offer various capabilities, the basic options include establishing a full-limit set that specifies milligram dose per time interval, a more limited set that might establish volume infusion rates only, or a label only (or a no limit) set for drugs such as cefazolin that can be safely pushed when necessary.

Building the Drug Library

Building the drug library can be a challenge. Do not underestimate the amount of time necessary to tailor the library to the needs of your specific patient populations. While there is currently no clearinghouse of drug libraries available, your vendor may be able to assist you in building a template, and reaching out to other institutions to discuss their library development process can be very helpful.

Essential to creating an effective drug library is having the pump in hand and using it when you are building or improving the libraries. By previewing exactly what the users will see, you can prevent flaws from being built into the system. Another key step is to understand your institution’s drug usage patterns. Create a list of all drugs that are normally infused and then review how they are used in different care settings. For example, we learned that our ICU quite often starts nicardipine infusions at a higher dose than our soft upper limit. We chose to retain the soft upper limit so as to alert the user that while they can start at this dosage, they must keep in mind where they are in the dosage range.

When building the library, avoid dosing conflicts within your own system. For example, we originally established an upper soft limit infusion rate for phenytoin at 25 mg/min with an upper hard limit of 50 mg/min at the direction of our ICU pharmacists. However, with our loading doses, we were using 1000 mg/100 mL with a label instructing the clinician to infuse it in 20-30 minutes, which exceeded our upper soft limit. This practice would conflict with the upper soft limit warning. To resolve this, we raised our upper soft limit to 40 mg/min. It is imperative that hospital drug dosing and usage patterns do not conflict with established safety limits. This causes general confusion, especially for nurses at the bedside. We have learned to research the literature; understand the drug usage patterns of our institution; and ask for input from our users, including nursing, critical care medicine, emergency medicine, oncology, pediatrics, and anesthesiology. Only then do we review and recheck all of our dosing data. Our library is a continuous quality improvement process that we work hard on because we know it prevents errors and saves lives.
{mospagebreak title=Page 2&heading=Page 2}
Monitoring the Drug Library
It should be pharmacy’s responsibility to continually monitor the library and make necessary revisions and additions. Such scrutiny will keep the library operating accurately and efficiently and build user confidence. The pumps offer a wonderful opportunity for on demand data collection. You can view how the pumps are being used throughout the institution or in individual patient care areas. With these pumps, every keystroke is recorded allowing you to review practice patterns and examine previous events in minute detail, which can sometimes reveal surprising data.

In addition to reviewing the reports for system inefficiencies, solicit recommendations about safety limits from all departments. All recommendations for changes to the drug library database should be heard whether they come from pharmacy, nursing, anesthesiology, or medicine. Ask clinicians to describe their frustrations and what they feel must be changed. In our hospital, the upper soft limits for the drug meropenem were originally set so low that every time it was used the limit was exceeded. If there is no good reason to keep the limits intact, then by all means change them. The open-ended question to clinicians should be “What do you see in our medication dosing parameters that should be changed?” We use our hospital’s Pharmacy and Therapeutics (P&T) Committee as the final authority on changes. In most cases, we worked the issue out ahead of time, and presented it to the P&T Committee not so much for input as for final validation. Nonetheless, it is key to establish procedures for updating the libraries.

We update the drug library periodically according to need—if we see a need to improve a situation in a particular CCA, then we make adjustments. We use a wireless system for the updates, which I strongly recommend. Using wireless communication, every pump that is plugged into a power outlet (not just running on battery), automatically receives the update. The pump notifies the user of the availability of a new library and the user can accept the update when the pump has finished an ongoing infusion (or at another convenient time). The update takes about five to seven minutes to complete. With a wired system, updates tend to be very infrequent because they are so cumbersome to download. While wireless updates are simple to accomplish, remember to alert nursing when an update is scheduled.

CQI Reports
Real-time reports and data gathering are available through the smart pump’s software, which provides advanced data transfer technology to create benchmarks, and monitor, evaluate, and improve the IV medication delivery process. The on demand reports are a vital component of patient safety solutions and identify therapeutic trends, drive quality improvement initiatives, and monitor and measure implemented changes. The reports, which can be scheduled and sent via email, provide the following:


Show compliance with safety software

  • Show clinicians’ responses to alertsn Review whether dosing limits are clinically appropriaten Identify current practice trends
  • Demonstrate maximization of pump functionalityn Identify operational improvement opportunitiesn Review pump utilization and drug library status
  •  Reports on asset tracking, asset utilization, and infusion status provide the physical facts of the pump: when and where was it last used and whether it is currently running.


There are clinical reports, such as those that present an infusion summary that provide very detailed information about each medication being administered (Fig 1). These drill down reports are crucial because they provide a summary of drug library utilization, including:

  • Frequency of programs and alertsn Breakdown of clinicians’ responses to alertsn Specifics of overrides and edits
  • Clinical safety and workflow efficiency implicationsn User response at the final confirmation delivery screen


Some of the most useful reports are those that provide override variance detail and edit variance detail as they communicate when an override or change in the entry for the hard and/or soft upper and/or lower limits of a medi­cation occurs. The reports detail the CCA, the medication, the date, and the time the alert occurred. Presently they are not interfaced with the patient medical record, but we hope to have this connection in the future.

Reading these reports is a study in improving the efficiency and safety of our drug administration practices. How many instances of alerts or actual edits of medication dosages are occurring? Are the soft limit alerts frequent enough, as well as founded in good pharmacology and therapeutics, to justify changing the drug library’s parameters for a particular medication? The goal is to avoid a situation where the limits are constantly overridden, as this will eventually become a habit, making the limits meaningless. We should change the drug library rather than allow regular overrides of the limits.

When we first implemented the pumps, we simply used the old TKO (to keep open) rate of 40 mL/hr as the lower soft limit for intravenous fluids (IVF). Because of the accuracy of these pumps and their ability to go down to single digit numbers, we discovered that we were frequently overriding the lower soft limit. So we changed the soft limit to 10 mL/hr in the drug library to keep all of us from getting in the habit of overriding an alert.

For antibiotics, we initially had one entry, whether the agent was cefazolin, piperacillin/tazobactam, meropenem, gentamicin, etc. We now have information on each specific agent (Table 1). Safety limits for a low-risk antibiotic such as cefazolin will differ from high-risk agents, and we wanted our nursing staff to become accustomed to using individual antibiotic names. Our infectious disease pharmacists, after viewing quite a few overrides with antimicrobials on the pump reports, decided to research some prolonged infusions versus conventional intermittent dosing. We changed our drug library after incorporating these findings in our normal continuous quality improvement (CQI) processes.

Inappropriate Safety Limits of Specific Medications
Phenytoin is a medication that should not be infused too quickly because the patient can experience pain as well as adverse events such as arrhythmias. With such medications, appropriate limits must be established. For phenytoin, our upper soft limit is 40 mL/min, and our upper hard is 50 mL/min. These limits are very well defined in the literature.

Another medication that should not be infused too quickly is the antibiotic, levofloxacin, so we need upper hard limits to avert adverse effects. However, levofloxacin is a medication that also needs lower soft limits because if it is infused too slowly over too long a period of time it never reaches maximum plasma concentrations and therefore the maximum antimicrobial effect is lost. See Table 2 for the safety limits for levofloxacin in our drug library.

With heparin, a very critical drug, we initially concentrated on the upper soft and hard limits, but neglected to establish soft and hard lower limits. Focusing only on the upper limits for heparin can lead to missing low dose errors that may result in therapeutic failures and patient harm. Reviewing our usage reports revealed problems with not setting these lower safety limits. For example, we found errors where the pump was set at “7” and the user assumed they were setting the pump at 7 mL/hr; in actuality they were setting the pump at 7 units/hr. To avoid this problem, we instituted a lower soft limit of 700 units/hr and a lower hard limit of 400 units/hr (Table 3).

We also found prescribing errors while monitoring infusion reports. When we looked at the override variance detail report, we saw that our neurology patients in the emergency department were being given a sub-therapeutic dose of heparin. After discussions and research, the dosing practices of heparin in this population were changed. The report helped us identify an opportunity to improve drug therapy in a critical care population.

Although standardization, when possible, is one of our goals, we still want our clinicians thinking about infusion rates. There are occasions when we choose not to change a limit that may be commonly overridden. For nicardipine, the override variance detail report showed us that the soft upper limit of 10 mg/hr programmed into our drug library was routinely overridden for 12 mg/hr infusions. After review by pharmacy, nursing, and medicine, we decided not to change the soft upper limit of 10, in spite of the many overrides (Table 4). We felt it was prudent for the user to be reminded with this drug, that the dosing was in the upper limit of the normal dosing range.

Critical Catches
We use propofol, a high-risk drug, on a routine basis. We consulted with anesthesiology to add an upper hard limit that they would not expect to exceed. What we came up with was an upper hard limit of 250 mcg/kg/min. Hopefully, the need never arises to infuse at that level, but if a clinician inadvertently enters 500 instead of 50 that error will be caught because of this upper hard limit. We strongly believe upper hard limits are life saving and should be in the drug library for all medications, even for those agents for which upper limits are difficult to determine.

For potassium infusion rates, the edit variance detail report showed a 200 mL/hr rate instead of the intended 20 mL/hr rate. This mistake was avoided because the drug library had an upper hard limit of 175 mL/hr. Adding or subtracting a zero is critical and can be caught through the use of upper hard limits (Table 5).

At our institution, we started with an overall pump compliance rate of 35% in all of our CCAs. In other words, the pumps were being used appropriately just over one third of the time. After one year, we achieved a 55% compliance rate, and in November 2008, at the two year mark, we achieved a compliance rate of 96.6% (Fig 2). This increase reflects the hard work of our staff in improving the libraries. With better libraries, nursing became more compliant and nurses encouraged each other to use the pumps. We publicized some of the critical catches to the staff and they were impressed. Our oncology center has always been the CCA with the highest compliance rate, partially because of the critical nature of the medications and the natural caution required for their delivery. In addition, the oncology pharmacists extended a significant effort determining each medication and its appropriate dosing, container volume and overfill.

Concentrating on high-risk medications, such as heparin, and stressing the safety features of having hard and soft limits for these drugs can increase compliance. With drill downs on individual medications, you can monitor all of the patients receiving that medication and identify trends. Individual CCAs can be audited and real-time feedback given to the nurse manager and staff about the findings and the importance of the pump and the drug libraries. The importance of feedback from the frontlines cannot be stressed enough. If you spend time rounding, you will understand the clinician’s problems and limitations when using the pump. We take our reports to a  monthly nursing directors’ meeting whose attendees include the directors of medical/surgery, women and children’s health, intensive care units, etc, and we discuss compliance and quality and how the two are linked. These reports are integrated into our current CQI processes and can result in medication standardization.

We have seen how the critical catches with an intelligent infusion system can be life saving. We can improve the quality of IV administration by establishing a degree of standardization, incorporating safety limits, and implementing optimum administration processes. The pump has made this much more possible.

We have received great value from our investment in smart pumps. Because these safety systems help catch potentially severe or even fatal medication errors at the bedside, they provide an excellent return on investment.

READ MORE: Continuous Commitment to Safety

Continuing Education Exam


Like what you've read? Please log in or create a free account to enjoy more of what has to offer.

Current Issue

Enter our Sweepstakes now for your chance to win the following prizes:

Just answer the following quick question for your chance to win:

To continue, you must either login or register: