Continuous Commitment to Safety

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

Click here to visit the ProCE website for this program

The 220-bed Women and Children’s Hospital in Buffalo, New York is part of Kaleida Health, a five hospital health care system. The majority of our beds are pediatric, including a 20-bed pediatric intensive care unit (PICU), which is also a regional trauma center. There is a 64-bed neonatal ICU, which admits more than 700 patients a year, 200 of whom are transported from other facilities in an eight county area surrounding Buffalo. We are also a designated high-risk obstetrics center with fetal maternal medicine consultations and a hotline, so we are concerned with adult as well as pediatric patients.

Our primary concern when dispensing intravenous (IV) medications to a predominantly pediatric population is consistent accuracy in weight-based dosing. To this end, we need drug delivery products that enhance patient safety, reduce IV-infusion errors, and increase clinician productivity. At the Women and Children’s Hospital we found all of these capabilities with an intelligent infusion pump, and we went live with these smart pumps in February 2008.

At the center of the pump’s safety system is the software, which offers clinicians programmable drug libraries specific to the clinical care areas (CCAs) or patient populations defined by each individual hospital. Our infusion device has the capacity for up to 40 different CCAs with up to 400 medications per care area. To ensure safety software compliance, the pump requires the selection of a CCA drug library when turned on and programmed.

Multidisciplinary Approach to Drug Library Development

The drug library development process requires a significant effort. Creating and revising the drug library is a continuous process, driven by input from the multidisciplinary team, pump report data, and other quality data, such as incident reports. The multidisciplinary team charged with creating our initial drug library was composed of pharmacists; nurses, including nurse educators, nurse managers, and transport team members; and physicians, including department chairs and anesthesiologists.

We found the input of our nurse educators to be particularly valuable as they shared not only their extensive clinical experience, but also their ability to create education programs that encourage clinicians to make the right decisions. All transport team members who occasionally use smart pumps during patient transport also required education on pump functions and drug library updates.

We also had some key physicians involved in the evolutionary process of the drug library. The co-chair of the hospital’s Pharmacy and Therapeutics Committee (P&T) is an attending physician in our PICU and a toxicologist, in other words a physician-pharmacologist. When disagreements arose as to dosing limits within the library, we would consult with the P&T committee or department chairs. Anesthesiologists are an important group to consider when creating a drug library. They are end users but, unlike other clinicians involved in the process, they are often the prescriber, dispenser, and administrator as well. Keep in mind that vendor consultants are a good source for tips of what has worked well in drug libraries in other
hospitals.

Increase Library Compliance

In the spring of 2007, the Institute for Safe Medication Practices (ISMP) issued a report highlighting how smart pumps are not smart on their own1. Emphasizing that drug library compliance is often very poor — as low as 30% to 50% — the report cited the following reasons for such low compliance rates:

  • Falsely low perceptions of risk
  • Failure to make timely corrections to the drug library for alerts that are not credible
  • Extra steps involved to use technology
  • Time pressures and clinical emergencies
  • Cultures supporting at-risk behaviors including technology work-arounds


After conducting extensive education during the implementation phase, we went to the front lines on rounds to see what actually prevents clinicians from using the infusion devices correctly. We found that one way to increase compliance was a very selective ordering of the institution’s medications within the library, starting first with the most common medications and their indications. Keeping the library current was also important to our staff, so we now update our drug library quarterly.

Another means to increase compliance is to make it easy to do right and hard to do wrong. With some pumps, when the device is initially turned on, the user has the option of using the pump with the drug library or in basic mode without the protections of the drug library. With other devices, once the pump is turned on the drug library is the default setting and it is somewhat difficult to opt out to use the basic mode.

Publicizing “good catches,” or situations in which the library’s hard and soft limits were lifesaving, is key. Let staff know what adverse events were prevented by using the pump. We want our staff thinking, “This could have been my patient, it could have been me that made that error.” Communicating the direct positive result of using the smart pumps to the staff will have a positive effect on the compliance rate.

Since we first started using the infusion devices ten months ago, our compliance rate with the drug library has ranged from 92% to 99%. Rather than becoming complacent, we are now looking at ways to achieve a 100% compliance rate (Fig. 2).

Creating a Drug List
Once the multidisciplinary team has been selected, formation of a drug list for the drug library is a necessary first step. Sources for creating a list include:

  • Continuous infusions
  • Standard concentration list
  • Stat log/fill list
  • Nursing requests
  • Sample libraries from vendors if available
  • IV medication policy
  • Intermittent doses


When creating a drug list, determining how to represent diverse patient populations within the same unit can be a challenge. We addressed this problem via creative naming and ordering. We also use TALLman lettering with look-alike sound-alike drugs. The importance of this approach is very evident when medications are listed alphabetically in the drug library.

Our medication order form requires all prescribers to include an indication for the prescribed agent, and this is particularly imperative for those drugs that have indication-dependent dosing parameters. In clinical practice, for example, the insulin used for pediatric diabetic ketoacidosis is different from that used to treat certain overdoses. Therefore, different soft and hard limits for each separate indication must be set in the drug library.

In a hospital with neonatal, pediatric, and adult patients, establishing varied — but limited — standard concentrations is key. In our hospital, each pump has a drop down menu listing available concentrations. Because not all concentrations are appropriate for all CCAs, it is important that your library match your standard of care and your policy on standard concentrations. For example, in our NICU the smart pump menu lists the most commonly used IV fluids/medications first, starting with total parenteral nutrition (TPN). If fentanyl is chosen in the NICU, the four approved concentrations will appear on the drop down menu. In the PICU, only two approved concentrations will appear (Table 1).
{mospagebreak title=Page 2 Continuous Commitment to Safety&heading=Page 2 Continous Commitment to Safety}
Standardized Dosing
Standardizing dosing units is another challenging task. Vasopressin provides a good example of these challenges as it has multiple dosage options and indications. Initially, we decided to include all the various dosing options in the library, so nursing would have a safe way to check the dose via the soft and hard limits no matter how it was
prescribed.

Our goal was to provide multiple dose checks per medication, including a rate check and a dose per kilogram check, while always maintaining a cap on total dose. Nursing voiced concerns that the multiple options for vasopressin made it confusing and difficult to choose the right indication. The simplification of the vasopressin dosages was on the agenda of a PICU quality assurance meeting that included the prescribers and nursing personnel. At that meeting, all involved decided upon one method for dosing vasopressin for each of its indications. With our next revision of the drug library, the nurses will see fewer options for this medication (Table 2).

With the listings of IV fluids in the pump, nursing actually requested more explicit and detailed options. They wanted each fluid to be specifically labeled to indicate whether it was a central line, a central line with heparin, an arterial line, etc. Nursing even requested multiple library entries for dextrose 5%.

We received feedback from anesthesiologists that the concentration they needed was not in the drug library. We could argue that they should be using standard infusions, but at the same time, we wanted them to have appropriate limits on each infused medication. In addition to listing the hospital approved standard concentrations, entries for anesthesia medications allowed input of the concentration (i.e.,___mg per ___mL). We have primarily set soft limits for anesthesiology, but soon we will have to commit to hard limits to avoid keystroke errors.

Increasing Usability
In our continuous customizing and refining of the drug library, our goal is always to make the pumps easier to set up, manage, and monitor and hopefully to achieve the ultimate goal of greater safety in the administration of IV medications throughout our hospital. While our smart pumps represent a quantum leap in terms of ease of use, we wanted our drug library to be user friendly as well.

One step in this process was to identify and minimize useless alerts. In our first library, we had an alert at the end of a syringe infusion, indicating there was still volume to infuse, which was a nuisance to the staff. The option for the delivery at the end of infusion was changed from KVO to “none” for medications given via a syringe. This setting was adjusted during the revision of the first drug library to eliminate the every two minute KVO Call Back alert. The limits on antibiotics were also adjusted to accommodate those (i.e. cefazolin) that may be delivered over 15 minutes (previous limits were based on 30-minute infusions).

The drug library must mirror the standard of care in your institution, not just in small overrides, but also in medication doses in general. We continue to require nurses to conduct a safe dose check prior to drug administration. The safe dose check is not replaced by the drug limits programmed into the library. The library’s soft limits have to encompass the majority of patients, but allowances must always be made for outliers titrated for specific patients and represent the hard limits programmed into the pump.

We customized our pump settings, such as alarms and brightness by CCAs. Our PICU nurses wanted the lights on all the time, while the floor nurses preferred to have the lights out at night. We also customized our report capabilities by CCA. We have five different pediatric floors, and all five have identical drug libraries. However, each library is identified separately because data is easier to combine than to separate. This allows the nurse educators and managers to learn what is going on in their unit and what educational points should be highlighted.

Additionally, we built a code CCA in the drug library. In an emergency, the code responder will need to access critical care medications, such as epinephrine and dopamine, which would not be in the general pediatric floor CCA library. Rather than scrolling through the entire ICU library and parsing through antibiotics and neuromuscular blockers, the medications are quickly available in the code drug library. It is also important to create a training CCA. While learning, nurses should hit hard and soft limits on purpose to see how the pump operates. With a training CCA, that data will not skew the actual patient data stored in the pump.

Our NICU and PICU chose different libraries and have different dosing parameters. In the NICU, patients are weighed in grams, in the PICU, in kilograms. Even though weight is entered in two different ways, the pump checks the dosage using kilograms and adjusts accordingly. The PICU enables the bolus function off the primary, whereas the NICU does not allow it. Other functions vary between these units, and the drug library is customized to accommodate such variations. (Table 3)

Avoiding Common Errors
Knowing what infusion errors occur at your site and why will allow you to design the drug library to avoid common errors. For example, we knew from incident reports that our TPN and intralipid rates were sometimes inadvertently interchanged. To address this we made the intralipids in NICU weight based with a daily maximum of 4 g/kg/day, although we still allow programming in mL/hr. Should the nurse enter the dose in mL/hr and accidentally put in the TPN rate, the pump would back calculate and give a soft and hard limit as well. In reviewing our pump reports, this error was avoided over 40 times in July and August alone, making for a nice return on investment (Table 4).

Because the PICU uses different concentrations of dopamine than the NICU, a challenge arose with concentration per milliliter versus total medication and total volume. We found that nurses were choosing 800 mcg/ mL because the dopamine bag was labeled 800 mg/ 250 mL. To resolve this we removed the two lower concentrations from the PICU library. Should a lower concentration be required, it is available in the NICU drug library, but must be obtained from the pharmacy.

Catches in the First Week
During our first week using the new pump, we were called to the PICU because of some difficulties with intravenous immunoglobulin (IVIG). We had our old pumps on standby in case anything was to go wrong. The new pump caught a decimal error. The ordered and programmed infusion rate was 0.8 mL/kg/min instead of the correct rate of 0.08 mL/kg/min.  Because the technology was new to us, the users assumed it was the pump, not the medication order that was incorrect. The order was corrected and the infusion was safely administered.

We also had a critical catch with insulin. Our library has an upper hard limit of 0.2 units/kg/hr for our pediatric diabetic ketoacidosis patients, although it is not inappropriate to start a younger patient at 0.05 units/kg/hr. One user accidentally programmed in 0.5 units/kg/hr, a decimal error that would have caused a tenfold overdose. We have also caught subtherapeutic doses of heparin and life-threatening high doses of remifentanil (Table 5).

Challenges of the Multi-campus Health System
We are a multi-campus health system with both adult and pediatric sites. Because the Women and Children’s Hospital also has adult patients, we need to have the same limits for the same patient populations across the health system. To achieve this, I have access to both the adult and pediatric libraries and the pharmacist who maintains the adult library also has this dual access.

Wireless Technology as an Imperative
Given the significant investment of time and energy required to develop and build a drug library, it is imperative that the library gets to every pump in a timely fashion. Using biomedical engineering to update each pump individually is incredibly labor intensive. Pushing the drug library wirelessly, on the other hand, is relatively effortless. In addition, with wireless technology, data is available in real time, allowing you to run daily reports of “no drug selected” entries, or query high-risk medication alerts, schedule the frequency of reports, etc.

Conclusion

During our first week with the new pump, we anxiously anticipated the success of the drug library. Now nurses are asking about adding medications that are not in the drug library. They are requesting tighter hard limits on certain medications, and they want to see the real-time data. They have discovered what a great tool these pumps can be in the total quality assurance process. They realize the value of the data available from these pumps to the ongoing improvement of patient care. n

Reference:

1. Institute for Safe Medication Practices. Smart pumps are not smart on their own. ISMP Medication Safety Alert! 2007 Apr 19;12(8):1-2.

Click here to take the Continuing Education Exam

Login

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

Current Issue