United States Pharmacopeia (USP) published the first set of enforceable standards for sterile compounding in 2004.1 Intended to ensure patient safety, USP <797> describes standards for proper preparation and storage of sterile products.2 More recently, USP <800> was published, with the intent of protecting employees by decreasing hazardous drug (HD) exposure.3 Both standards have elevated the practice of sterile compounding by ensuring safe practices, yet neither addresses compounding accuracy or standardized dispensing of sterile products, resulting in a lack of direction for selecting product preparation and verification methods. The Institute for Safe Medication Practices (ISMP) has identified independent verification of components, prior to addition into the final container, as a best practice.4 The recommendation endorses utilizing technology to assist in verification and suggests that the use of bar code scanning of base fluids and ingredients should be considered as a minimum requirement. With the availability of gravimetric-based IV workflow software, individual components can be verified to ensure accuracy and increase the safety of IV preparations.
The University of North Carolina Medical Center (UNCMC) is a public, not-for-profit, academic medical center with more than 800 beds and a 75-bed community hospital, which serves as an extension of the medical center. With three onsite and five offsite infusion centers, UNCMC has over 150 infusion chairs. The Cancer Hospital infusion/inpatient pharmacy is UNCMC’s largest infusion pharmacy, dispensing an average of 215 patient-specific sterile products per day for adult and pediatric patients across inpatient and outpatient settings. The five offsite infusion centers comprise three oncology and two non-oncology infusion sites, with each site preparing between 10 and 85 patient-specific sterile products per day. In an effort to increase safety in the IV workflow process, UNCMC began investigating the possibility of implementing gravimetric-based IV workflow software almost 10 years ago.
Retiring the Volumetric Proxy Verification Method
Prior to implementing IV workflow software at UNCMC, IV admixture components were verified through manual inspection of the volumetric preparation method. In the infusion pharmacies, a proxy verification method was utilized, wherein the pharmacy technician pulled back the syringe containing the drug to the appropriate demarcation, used a permanent marker to indicate on the syringe the level to which the medication was measured, and then injected the drug into the final container. The pharmacist verified the product preparation by inspecting the empty syringe and visually verifying the permanent marker line. ISMP has reported errors causing serious harm and death associated with the syringe pull-back method.4 Because pharmacists are unable to determine if the syringe reflects the accurate amount added, or if a pulled-back syringe is partnered with the correct vial of medication, ISMP recommends this method not be used.4
UNCMC conducted a study to determine the accuracy of the volumetric preparation method for chemotherapy doses. Of the 1156 doses included, the mean percent volume difference was -0.53%, with a range of -64.9% to 94.2%.5 The high level of variability among the chemotherapy doses prepared via the volumetric method provided justification for the subsequent purchase of gravimetric-based IV workflow software in December 2015.
Implementing Gravimetric-Based IV Workflow Software
The exploration of IV workflow products began in 2009, although the project was postponed, as the process of changing from one electronic health record (EHR) to another took priority shortly thereafter. Once the final EHR was selected in 2013, choosing gravimetric-based IV workflow software again took precedence.
The pharmacy department sought a feasible solution for both HDs and non-HDs that could be implemented at each sterile preparation site. To assist in an objective review, a scorecard was developed for each product under consideration (see FIGURE 1). The metrics were numbered arbitrarily, which allowed for an unbiased assessment. For each metric, the availability, utility, and ease of operation were assessed through a functionality scale from zero to five (zero=no functionality; five=functional, stable, and user-friendly). The pharmacy department determined a global priority level for each function, which applied across all products.
Mandatory functionalities included gravimetric capabilities and embedded hard stops within the software. The goal was to have hard stop functionality during technician preparation to facilitate error identification in real time and also preclude the ability to proceed without correcting the error. This included hard stops within the bar code scanning process and for out-of-tolerance-range alerts during dose preparation. Critically important was the product’s ability to ensure accuracy within a set tolerance range, prior to reaching the pharmacist for the final product check.
Aside from product functionality, cost and the vendor’s quality of service were additional considerations during product selection. The diversity among our planned implementation sites complicated the determination of optimal pricing. Vendors who priced based on the volume of products prepared through the system were advantageous for smaller, low-volume locations, while flat-rate pricing per workstation benefited larger volume sites. Ultimately, we determined that the flat-rate pricing model was most cost-effective based on the high volumes prepared at the larger sites.
Vendors were brought onsite to demonstrate their systems to a combination of pharmacy leadership, pharmacists, and technicians. The selection was narrowed down to two vendors; both gave product demonstrations to the directors of pharmacy roundtable (which includes every director of pharmacy in the UNC health care system). Based on feedback from each review, UNC health care pharmacy services selected a solution which provides gravimetric dose verification integrated with bar code verification and electronic documentation. The software guides technicians step-by-step through sterile product preparation and includes gravimetric measurement hard stops. Pharmacist verification is streamlined using stepwise visual images combined with a detailed preparation protocol that includes the time of preparation, identification of the preparing technician, and gravimetric measurement data.
As the first site in the health system to go live with the new EHR, UNCMC was selected to be the first entity to install the new IV workflow software, while the other facilities focused on the EHR rollout. The UNCMC cancer hospital was the first HD preparation location to go live.
The timeline from product selection to implementation is outlined in FIGURE 2. Contract negotiations took place for over 1 year, mainly due to changes in the IT division, which included new committees and approval processes. The implementation kick-off meeting occurred 1 month after contract execution. The pharmacy department determined who, in addition to the pharmacy implementation team, would have administrative access to the system. Administrative users serving on the pharmacy implementation team included the pharmacy clinical manager, the lead pharmacist, the lead technician, and select pharmacists within the oncology department. Colleagues within the IT department identified point persons to guide the integration process between the EHR and IV workflow software.
The pharmacy implementation team’s areas of focus were building the drug database library, validating the database build integration with interfaced EHR orders, validating the IV workflow software process, and streamlining workflow. The HD database was built in its entirety by UNCMC over a 4-month period. Obtaining information regarding beyond-use dates (BUDs), stability information, fluid compatibilities, active ingredient amounts, and product densities occurred primarily through the institutional HD compendium. When information was not readily available, pharmacy students on administrative rotations and UNCMC’s Drug Information Center contacted manufacturers and conducted literature searches to retrieve the data. Building the entire database up front and then testing the database with the EHR interface was the initial goal; however, during the process it became clear that phased testing would be more appropriate. Ultimately, phased testing consisted of building a few drugs with similar preparations, testing, and then modifying the drug builds as necessary prior to building out all similar preparations.
Proper education and staff training aided in successful implementation. Targeted super-users were trained with the goal of then training fellow staff members. The vendor led 2 days of onsite system setup and preparation training, as well as 2 days of 2-hour breakout sessions for individual super-users. Individual training sessions included a system overview, workflow details, and hands-on application. Education of end users occurred in two phases: an introductory phase and a hands-on application phase. All staff completed online training modules developed by the vendor prior to participating in hands-on training. All technicians and all super-user pharmacists participated in mock order preparation using two different formulations of low-cost drugs (eg, ampicillin powder for injection and fluorouracil solution for injection). Orders were entered in the test environment of the EHR, and the test orders interfaced with the gravimetric-based IV workflow software test environment to simulate a live environment. Each end user pharmacist verified test-prepared products in the software testing environment prior to go-live.
Software and Logistics
Prior to setting up the cleanroom with the new technology, the implementation team visited another institution that was live with the system. Based on information gleaned from this visit, IV workflow software monitors were placed within the hood and the hardware setup was managed as a collaborative effort between the IT department and the vendor. The vendor came onsite to set up all hood units and served as a resource throughout the implementation process. UNCMC worked with an external cleanroom certification vendor to reinforce the sidewall of the hood, install mounting arms for the monitors, and identify a method for system cables to pass out of the hood for all necessary utility connections.
The IT department examined the admission, discharge, and transfer (ADT) interface to ensure alignment between the EHR and the IV workflow software. Interface coordination guarantees that orders discontinued in the EHR result in discontinuation of orders within the IV workflow software queue. In addition, patient locations are updated following transfers to ensure accurate location information on drug labels. At this time, information exchange is limited to EHR orders propagating into a workflow queue within the IV workflow software. The pharmacy department tested the unidirectional interface by entering orders in the EHR’s testing environment and confirmed receipt by the IV workflow software. A bidirectional interface is now available and is currently being tested, with a planned implementation this year.
The initial strategy for implementation was the “big bang” approach, wherein all HDs would be implemented into the IV workflow software at once. However, the specificity of the interface to the EHR build interfered with some orders crossing through the system. Thus, it was decided that phased testing should occur to ensure all orders built would interface correctly. To avoid delaying implementation and to provide appropriate time for the build team, the IT department, and the vendor to work through these interface-related issues, it was decided that a phased approach for implementation would be prudent. This would also ensure that production times within our health system’s large infusion center could be maintained.
Prior to the go-live date, the pharmacy implementation team and the vendor were in touch on a weekly basis. Four hazardous medications were implemented in August 2016: docetaxel, oxaliplatin, paclitaxel, and pemetrexed. These medications were chosen because they are frequently used but require different preparation steps, which provided an opportunity for staff to gain experience with the system. Two weeks later, six additional drugs were added, followed by 10 more at a later date. Currently, approximately 70% of all HDs are dispensed through the gravimetric-based IV workflow software within the UNC cancer hospital. Two of the five offsite HD preparation sites have now gone live, with the remaining sites awaiting USP <800>-compliant renovations prior to scheduling their implementation.
Gravimetric verification achieves the ISMP best practice recommendation and improves the safety of compounded sterile products.4 Every production step is documented through the gravimetric functionality, and the hybrid model purchased by UNCMC includes a camera that records high-definition photos. At verification, pharmacists can confidently determine that the selected product is correct, as is the dose drawn into the syringe and the volume injected into the bag.
With narrow therapeutic range drugs, such as chemotherapy, dose accuracy is paramount. Gravimetric verification ensures the patient is neither over- or under-dosed. Built-in hard stops during bar code scanning and technician preparation (for products out of the accepted error range), provide real-time feedback to the technician. For example, the system caught a potential error with a manufacturer-assistance replacement stock vial. The manufacturer replacement program provided docetaxel 20 mg/mL vials, although docetaxel 10 mg/mL is the standard UNCMC concentration. The new concentration was added into the inventory and placed in the 10 mg/mL storage location, and the technician pulled the higher concentration vial for preparation. Upon bar code scanning, the IV workflow software recognized this as a new product and required the new strength be added to the drug database. While preparing to draw up the dose, the technician noticed that the volume the IV workflow software was requiring was half that of the necessary volume populated on the patient-specific drug label. The technician consulted with the pharmacist, at which time the potential error was identified. While it is possible that this error could have been caught manually, it would have been difficult to identify. The gravimetric-based IV workflow software prevented dispensing and administration of double the ordered dosage.
In addition to enhancing safety, gravimetric-based IV workflow software can track waste and improve standardization. The system automates tracking of vial BUDs, as well as milligrams left in partial vials. This helps reduce waste by ensuring that vials with the nearest BUDs are used for preparation first. Furthermore, gravimetric-based IV workflow software promotes product and workflow standardization using drug-specific protocols; all technicians follow identical preparation procedures, which are captured within the system.
Challenges are inevitable when implementing any new technology. During the formulary build, determining product densities presented a challenge. Accurate density information is crucial to establishing gravimetric accuracy; obtaining reliable data required a resource-intensive search. Fortunately, density information for chemotherapy is generally more available than for other drugs, so we were able to compile the required data for HDs. The database requires continuous upkeep; hence, a standardized method of accessing density information would be valuable. Our pharmacy department is currently researching the best method for determining densities not published by the manufacturer.
The most significant and time-consuming challenge to overcome stemmed from the dependence of the IV workflow software’s functionality on the institution’s specific EHR build. Differences among EHR builds inhibit the use of a universally integrated database; thus, any customization of the IV workflow software to an institution’s EHR is required for all systems. Additional challenges associated with the EHR interface were products requiring a commercially unavailable volume, commercial bags that contain overfill, and split syringes. At UNCMC, the production of products requiring a commercially unavailable volume is managed on an automatic compounder. However, the system was unable to scan and identify the patient-specific compounded bag, which halted workflow. New logic in the gravimetric-based IV workflow database had to be created to allow this functionality to occur. Another challenge we encountered arose from overfill within commercially available fluid bags. As an institutional standard, production labels include overfill volume as an additional line item. The technology interpreted overfill as an additional ingredient to be injected into the bag. Thus, new logic was created to bypass this step without the removal of overfill from the production label. Finally, preparation of spilt syringes is an ongoing challenge, and new logic in the IV workflow software is being evaluated to provide a resolution.
Next steps include integration of all products with the gravimetric-based IV workflow software. Following the final phase, we expect to prepare approximately 90% of all HDs using the system. Gravimetric preparation may not be feasible for all products. Consider investigational drugs, for example; until institutions require sponsors to provide density information and bar codes on investigational drugs prior to agreeing to perform the study, they cannot take full advantage of the safety features in the software. Currently, the medication may be entered in the system as a volumetric preparation to take advantage of bar code scanning and visual photo documentation.
The upcoming bidirectional integration between the EHR and the gravimetric-based IV workflow software will assist in ensuring the exact NDCs used are captured. The IV workflow software will also integrate with dose-tracking software built into the EHR to allow nurses and pharmacists access to real-time preparation status.
Patient safety and ensuring accurate chemotherapy preparations motivated our exploration of gravimetric-based IV workflow software. In addition, implementing gravimetric-based IV workflow software meets ISMP’s best practice recommendations for chemotherapy preparation through integrating bar code scanning and gravimetric preparation technique. The gravimetric-based IV workflow software has improved safety of our drug preparation process and meets the needs of our pharmacy department.
Charlotte Wells, PharmD, is a health-system pharmacy administration resident at the University of North Carolina Medical Center (UNCMC) in Chapel Hill. She is enrolled in the Master’s degree program at the UNC Eshelman School of Pharmacy. Charlotte obtained her Doctor of Pharmacy degree from the UNC Eshelman School of Pharmacy.
Nathan E. Barnes, PharmD, is the lead pharmacist for cancer hospital operations at UNCMC. He received both his Bachelor of Science and Doctor of Pharmacy degrees from the UNC Eshelman School of Pharmacy in Chapel Hill. His practice interests focus on high-quality, innovative approaches to cancer care.
Lindsey B. Amerine, PharmD, MS, BCPS, is the associate director of pharmacy for infusion services, investigational drug services, the medication assistance program, and revenue integrity at UNCMC. She is an associate professor of clinical education for the division of practice advancement and clinical education at the UNC Eshelman School of Pharmacy in Chapel Hill, North Carolina. Lindsey earned her Doctor of Pharmacy from the University of Wyoming School of Pharmacy and her MS with an emphasis in health-system pharmacy administration from the UNC Eshelman School of Pharmacy while completing a 2-year health system pharmacy administration residency at UNC Hospitals.
Charlotte Wells and Nathan E. Barnes report no potential conflicts of interest. Lindsey B. Amerine served as a speaker for Becton, Dickinson and Company.
Enter our Sweepstakes now for your chance to win the following prizes:
Just answer the following quick question for your chance to win:
Entries are limited to one entry per person in any active sweepstakes.
Thank you for your entry.