Last year’s deaths and injuries associated with the distribution of contaminated methylprednisolone acetate syringes have prompted a great deal of uncertainty regarding future practice standards for compounding sterile products. Unfortunately, that tragedy is not the only incident that has contributed to the renewed scrutiny of sterile compounding practices. But due to its magnitude, those failures to ensure product integrity may become the watershed event that will ultimately change regulatory requirements and clearly define standards of practice related to sterile compounding. There are strong indications that in the near future commercial compounding centers may be restricted to compounding sterile products pursuant to a patient’s prescription or under FDA approval.1
Pharmacy Purchasing & Products’ April 2013 State of Pharmacy Compounding survey reported that 60% of pharmacy directors rely on sterile product compounding centers to source at least some of their sterile products.2 The items most commonly outsourced to compounding pharmacies include PCAs, pitocin/oxytocin, and epidurals. Furthermore, approximately 75% of hospital pharmacies have used compounding services in response to a specific product shortage. Financial challenges also have caused some hospitals to reduce pharmacy staff, further emphasizing the need to find safe, alternative methods of preparing compounded sterile products (CSPs). In this environment, more hospitals are considering the acquisition of compounding robots to produce CSPs in-house.
SUNY Upstate University Health System is an academic medical center in Syracuse, New York, comprising University Hospital, Golisano Children’s Hospital, and Community General Hospital. In total, there are 720 licensed acute care beds located on two campuses. The main campus is a Level 1 trauma center that treats both children and adults with complex illnesses and serious injuries. In 2009, prior to the expansion of our pediatric services and the acquisition of Community General Hospital, it became apparent that we would not have the facility and staff resources required to meet the increased demand for CSPs. Thus, as our IV production workload increased, we began to outsource approximately 7,000 CSPs per month to outside vendors. In 2010, we began to consider the purchase of an IV robot as an alternative to outsourcing.
Evaluating the Option of IV Robotics
To determine the viability of IV compounding robots for our hospital, we first had to define our goals, perform a financial analysis, and develop a business plan. The heightened demand for CSPs was due not only to the increased number of licensed beds, but also the implementation of a bedside bar coding program. The bar coding software automatically classified all IV admixtures as patient-specific and required a patient label with a bar code identifying each CSP individually, which required additional time. When we began to look at ways to manage these challenges, we took note of the accuracy and safety benefits of robotic technology, as well as the fact that we were paying a premium to outsource CSPs, and these factors made robotic IV technology an attractive option. In response, we developed a financial model to determine if the potential savings associated with in-house, robotic CSP production would offset the cost of purchasing an IV robot. Initially, we were unsure if the savings would be significant enough to justify the investment in the robot, but after completing our analysis, we were convinced that the return on investment (ROI) would be substantial.
As early adopters of robotic IV technology, we did not have access to an established practice model for selecting a robot. Therefore, we developed our own set of evaluation criteria, which included the following:
Additional considerations, beyond the technology, included:
At the time we began evaluating this technology, there were four IV robots available and each presented unique advantages and limitations. For example, IntelliFill i.v., manufactured by ForHealth Technologies (acquired by Baxa, now a part of Baxter), offers effective throughput for small volume parenteral products repackaged in syringes, but does not prepare IV bags. The CytoCare robot, manufactured for hazardous drug compounding by Health Robotics, was removed from the US market early in 2011. Having ruled out those options based on our specific needs, we focused on the two technologies we felt would fulfill our requirements—Health Robotics’ i.v. Station and Intelligent Hospital Systems’ RIVA (Robotic IV Automation).
Robot Selection Criteria
IV robots are precisely calibrated to meet specific tolerances, such as how deeply to insert a needle into the injection port of an IV bag. But in order to program a robot to perform this function accurately, one must first measure the exact dimensions of the product container and then manipulate the robot with geometric programming to repeat the exact movement each time it recognizes the product container using bar code identification. The combination of bar code identification and geometric validation adds an element of safety that is not obtainable using manual processes. Robot movements are so precise that a vial stopper can be punctured multiple times through the same penetration site without causing coring, stopper permeability, or leaking.
Robotic technology requires not only precise initial calibration, but frequent testing and recalibration to ensure it functions properly and stays within established tolerances—procedures that can best be accomplished by the manufacturer’s trained field service representative. The vendor’s technician should perform testing, calibration, and preventive maintenance on a regular basis (eg, weekly, monthly, every six months, and annually, as required) to prevent any variations in functionality and performance. Accordingly, it is important to identify what your service and maintenance program includes prior to leasing or purchasing an IV robot.
In addition, IV robots utilize closed environments similar to isolation chambers, maintaining ISO 5 air quality throughout the compounding process. Both RIVA and i.v. Station use UV light to further reduce the possibility of product contamination. Combined with robust robot cleaning procedures and decontamination steps performed by pharmacy technicians, such as alcohol swabbing of vial stoppers and injection ports, these processes promote product safety.
Keep in mind, the compounding process must be controlled by a licensed professional, so pharmacists should be the only staff members to take product measurements and input the necessary geometric parameters, NDC numbers, and specific gravity information into the drug library software.
Batch vs Patient-specific Compounding
Our IV robot analysis occurred at the same time our organization was considering a transition to a new hospital information system (HIS). So we knew that if we built an interface between a new IV robot and our existing pharmacy system, we would need to build a new one when the next HIS was implemented. This reinforced our decision not to use the robot for patient-specific dosing. By configuring the robot to do batch mixing, we did not need to spend time developing a patient-specific interface both before and after the transition to the new HIS. In addition, it was easier to calculate ROI by comparing in-house batch production to the cost of outsourcing CSPs to an IV compounding pharmacy. If you are planning to use the robot for patient-specific CSPs, determine the time and cost required for the robot vendor and pharmacy software vendor to build the interface. Depending on the interface requirements and pharmacy system compatibility, you may be inclined to choose one robot technology over another. Both i.v. Station and RIVA can use standard HL7 interface software.
In addition, patient-specific compounding may impact throughput. Consider what products will be prepared and how the queues will be configured. Better throughputs may result if small batches of the same product are run at designated times rather than preparing individual doses one at a time on an as-needed basis. Both RIVA and i.v. Station can prepare syringes from 1 mL to 60 mL and IV bags from 50 mL to 1,000 mL, and RIVA is also capable of preparing 25 mL IV bags. Our hospital uses the robot to prepare both bags and syringes. We also adjust our batch sizes based on changes in product usage and the expiration dates that can be assigned in order to reduce waste.
Hazardous vs Nonhazardous Compounding
The decision to use the robot for hazardous or nonhazardous IV production will influence the robot’s installation placement and the cleanroom configuration. i.v. Station is designed for nonhazardous compounding and its chamber utilizes positive pressure, while the i.v. Station Onco is designed with a negative pressure chamber for hazardous drug compounding. As such, both i.v. Station, and the more recent model, i.v. Station Onco, are designed with single ISO 5 chambers within the enclosed environment of the robot. RIVA is designed with two ISO 5 vestibules to control air quality at the locations where supplies are loaded into the robot, and the compounding area itself is a self-contained ISO 5 area. RIVA can be configured onsite to meet either nonhazardous production or hazardous production specifications.
Even with these environmental control conditions, some hospitals are proactively installing hazardous compounding robots in ISO 7 (or higher) cleanrooms to be certain that air quality and safety requirements are satisfied. Pressure differential ratios and appropriate exhaust filtration equipment also are required to ensure appropriate containment of hazardous drugs.
When preparing production queues, use the largest vial size available; this will minimize the number of times the robot has to return to the supply location to obtain an additional vial, thus reducing production time. Determining average usage rates for each product and the amount of time required to complete one batch will aid in the development of production schedules.
RIVA can provide run times in advance, which aids in efficient scheduling. Check for changes in product utilization on the day of the batch run in order to minimize waste. A simple chart can be used to schedule production runs (see Table 1). At our hospital, adjustments are made daily based on changes in utilization and critical needs associated with drug shortages.
Product availability and internal carousel capacity will impact production as well. Keep in mind that the exact products for which the robot has been programmed must be utilized; any variation in product or container size will cause the robot to reject the queue. Thus, thoughtful and careful programming of the robot and development of efficient production schedules is essential and may require a significant time commitment. Properly building the queues will prevent geometric variations, which will, in turn, facilitate throughput goals and enable a smooth workflow. Initially, we reviewed approximately 100 CSPs for possible robotic production, including drugs we were outsourcing, frozen premixed IVs, and other standardized adult and pediatric drugs that were manually prepared in batches to meet patient needs. Of these, seven drugs suitable for robotic production were identified.
Each robot manufacturer provides specific guidelines on the full range of dose measurements and vial size capacities, and both i.v. Station and RIVA are compatible with many available brands of syringes; nevertheless, it is still prudent to verify compatibility early in the evaluation process. Each syringe will require a specific needle type, length, and gauge specified by the equipment manufacturer. Syringes can be purchased separately or with needles already attached. If syringes are purchased without needles, be sure staff is available to attach the required needles in an ISO 5 environment in order to maintain sterility prior to loading them into the robot. Other operational points to consider include:
If you plan to prepare both IV bags and syringes with the same robot, be sure the robot is capable of sufficiently labeling both. Also, confirm the labeling software can produce the type of bar code required to ensure consistency with your pharmacy system bar code format and point-of-care technology. With the growth in BCMA adoption, it is increasingly important to limit the number of bar code variations on the final product in order to avoid scanning failures. In addition, make sure the labels you use are commercially available from more than one vendor and that there are no concerns with ink smearing or absorption of the label adhesive chemicals into the CSP. Robotic system suppliers should be able to provide guidance in these areas.
Physical Dimension and Engineering Considerations
Two of the most obvious differences between RIVA and i.v. Station are the size and weight of each robot, which are important when determining where to locate them. Notably, RIVA is more than twice the size and weight of i.v. Station. To negotiate this, be sure to determine the floor load capacity, the space available in the pharmacy, and the workflow requirements before deciding which robot to acquire and where to install it. If sufficient space is available for the larger robot, then the increased storage capacity it provides may influence your choice.
Another key consideration for robot placement is accurately evaluating the HVAC requirements for maintaining proper cleanroom temperature during the time the robot is in use. This was a challenge for our design team, resulting in a dramatic underestimation of the cooling capacity required to offset the heat generated by the robot. Although the room had been designed to meet the robot’s tolerance of approximately 82°F, our team had not taken into account three key factors: The body heat generated by cleanroom staff, some IV solutions and other drugs must be stored at temperatures below 78°F, and forcing air through HEPA filters heats that air. Ultimately, our cleanroom had to be retrofitted with additional air conditioning units, at considerable cost, in order to maintain a buffer area temperature of 66˚F. Thus, it is vital to investigate and account for all possible heat sources to ensure the cleanroom remains within the correct temperature during IV robotic production.
Training and Credentialing Staff
When designating staff to support the IV robot, keep in mind that the skill set is different than what is typically required for manual IV compounding. Although a thorough knowledge of aseptic technique and compounding practices is a prerequisite, an understanding of geometric programming, automation, and technology is also required. As with most staffing decisions, it is important to select individuals who display an interest and an aptitude for the role. It also is important to develop a dedicated team of pharmacists and technicians who receive special training and gain a thorough understanding of the robot’s functionality and limitations. These super users can then assist in training new employees when the robot’s production hours are increased.
Staff training at our hospital began with vendor-supported education, and took place over a three-day period. In-house training continued for many months afterward, as the team became familiar with the nuances of the robot and uncovered methods for streamlining production and maximizing throughput. Although a great deal of information can be gleaned from early adopters and user groups, IV robot utilization is still a relatively new field, so we continue to expand our understanding as we increase our drug library and production capacity. Some of the basic training modules that we use include:
In addition to the initial vendor training, we require in-house validation of our employees to demonstrate their skills in each of the above areas annually. Pharmacy staff is credentialed to work with our robot only after they have successfully completed the didactic training and a written exam, and have demonstrated their skills to the satisfaction of our senior sterile process pharmacist. The exam we utilize was created by our vendor, and we made some necessary modifications to ensure it properly evaluates the skills of our employees.
Support Staff Requirements
The number and type of production queues necessary for your compounding operation should dictate technician staffing requirements. Once the medications in the production queues are identified, pharmacy technicians enter the drugs, diluents, solution lot numbers, and expiration dates into the batch record for tracking purposes, a process that takes approximately 10 to 15 minutes. It is important to use products with the same lot number in a single batch so that you will know what CSPs to remove from active inventory should a product recall or an indication of final product contamination occur. The robot assists with this process by assigning a unique sequence number for each production queue; both i.v. Station and RIVA have this capacity.
Pharmacy technicians also are responsible for wiping down all drug and diluent containers with alcohol and removing IV bag overwraps and vial caps prior to loading the IV robot. Syringes, needles, and syringe caps are loaded into specially configured carousels, and once the inventory is loaded, each injection port is swabbed with sterile alcohol. The inventory preparation and loading process takes approximately 20 to 30 minutes. In total, it takes about 45 to 60 minutes to set up a queue and initialize a batch.
The robot’s ISO 5 vestibule area is configured with a horizontal laminar hood so technicians can perform other work while the robot is running. In addition to robot production duties, our technicians attach needles to syringes, prefill IV sets for chemotherapy infusions, and prepare other batch CSPs not easily adaptable to IV robot production. Each technician works a seven day on, seven day off schedule and spends approximately 6 hours per day involved in IV robot activities. If possible, incorporate dedicated staff for IV robot production in your initial cost estimates to help refine cost and ROI projections.
Two technicians are assigned to IV robot production and one pharmacist oversees the program during production periods. Once the robot is programmed and the production queues are set, the pharmacist’s responsibilities are limited to verifying that the technician is following procedures and that appropriate samples are checked to verify accuracy and sterility. The pharmacist reviews patient orders and inventory levels, and develops a daily robot production schedule. The pharmacist also is tasked with remotely monitoring the preparation process via video cameras; the pharmacist and technician communicate verbally using a portable communication device. This method allows us to maximize productivity by freeing our pharmacists to participate in other clinical activities while still ensuring appropriate compounding oversight.
Return on Investment
Ultimately, we selected a RIVA robot because we felt the inventory storage capacity, the comprehensive service agreement, and the support available from the vendor’s scientific advisory board were the best fit for our organization. We are pleased with RIVA’s throughput, although there may be an advantage in buying two i.v. Stations to achieve a higher throughput for about the same price as one RIVA.
We initially estimated it would take approximately three to four years to reach a breakeven point based on the purchase cost and the necessary cleanroom renovation costs. However, later it became apparent that our actual renovation costs would be higher than originally expected—more than $150,000 more—due to design errors and the additional HVAC requirements. Moreover, there were costs associated with process validation and stability testing by an independent laboratory that we did not identify in our original analysis, totaling approximately $35,000. Likewise, we underestimated the time it would take to achieve a full production schedule, and so our first year’s savings did not meet our original saving projections of $200,000. To mitigate these issues, be sure to include all costs—acquisition of the robot, renovations, HVAC updates, and process validation and stability testing—into the ROI analysis to achieve accurate financial predictions. Now that we have completed our first year of robotic production, we have adjusted our breakeven point to four years, which is still an acceptable ROI. Table 2 illustrates the actual savings during our first year of IV robot production.
Because CSPs lack an FDA finding of safety, efficacy, and manufacturing quality, they are not interchangeable with FDA-approved medications. If an FDA-approved drug equivalent is available, that product should be obtained and used. If a CSP must be used, IV robot technology is ideally suited to achieve the uniformity needed for statistically significant USP <71> testing, as each dose is produced identically.
Direct human interaction in the compounding process is the leading source of potential contamination, thus IV robots offer engineering controls and closed environments that virtually eliminate this factor and provide product protection from external contaminants during the compounding process. The combination of bar code recognition and volumetric, geometric, and gravimetric measurements confirms a level of accuracy and safety beyond what is achievable with manual processes.
Ultimately, the decision of whether to switch to a compounding robot depends on the specific needs of an organization and must incorporate a balance between quality, safety, and affordability, as well as the possible impact of any future regulatory changes on the costs and benefits of IV robotics. Given the fluctuations in CSP availability and heightened scrutiny by regulatory agencies, IV robots will likely play an important role in the future of compounding centers and hospital pharmacies.
Comments in this article regarding practice standards are based solely on the author’s opinions and personal interpretations of existing guidelines and are not intended to be a legal interpretation of regulatory requirements. Guidance from both state and federal regulatory agencies should be carefully considered before making decisions regarding appropriate methods for validating the integrity of compounded IV products.
Steven J. Ciullo, BPharm, MS, MPS, received his BS and MS from St. John’s University School of Pharmacy, his masters degree in professional studies in health care administration from Long Island University, and completed an ASHP Residency at Mercy Hospital. He is currently the director of pharmacy services at Upstate University Hospital, Upstate Medical University in Syracuse, New York.
James Zahra, BPharm, received his BS in pharmacy from the University of Buffalo, School of Pharmacy and completed an ASHP Residency at EJ Meyer Memorial Hospital. He is currently Senior Manager for Compounded Sterile Products and IV Robotics at Upstate University Hospital, Upstate Medical University in Syracuse, New York.
Irina Pustovalova, PharmD, BCPS, received her doctor of pharmacy degree from the Ernest Mario School of Pharmacy at Rutgers University and completed an ASHP Residency at NYU Langone Medical Center. Irina joined the pharmacy department at Upstate University Hospital as a clinical staff pharmacist and also assists with IV robotic production. The authors would like to acknowledge the following pharmacy employees for their contributions to our IV robot service and this article: Heather Aldrich, CPhT, Pharmacy Assistant; and Samuel Kumar Velpula, CPhT, Pharmacy Applications Specialist.
The authors would like to acknowledge the following pharmacy employees for their contributions to our IV robot service and this article: Heather Aldrich, CPhT, Pharmacy Assistant; and Samuel Kumar Velpula, CPhT, Pharmacy Applications Specialist.
Ensuring Safe CSPs
Last year’s compounding tragedy, in which 749 patients were infected and 63 died after being infected with fungus from contaminated steroid injections,1 is one of the most serious incidents of patient harm from improperly compounded CSPs to date; however, similar events have occurred frequently over the past 15 years (see Table 3). To best ensure safety, the FDA recommends that when an FDA-approved drug is commercially available, practitioners prescribe that drug rather than a compounded drug, unless the prescriber has determined that a compounded product is necessary for the particular patient and would provide a significant medical difference for the patient compared with the FDA-approved, commercially available drug product.2 In light of this guidance and considering the heightened regulatory scrutiny since the compounding tragedy, pharmacists must carefully consider whether to prepare a CSP in the pharmacy or purchase a commercially available alternative. This decision should not be based solely on cost and should take into account patient safety concerns.
After testing numerous CSPs for contamination in 2001, the FDA concluded that 34% of those tested had deficiencies.3 In 2004, USP <797> emerged as the first enforceable, official set of standards meant to guide the preparation of sterile products; however, not every state adopted the standards. The variation in practice from state to state created uncertainty regarding the requirements, but incidents such as the recent compounding tragedies may serve as the impetus for ensuring that regulatory requirements are clearly delineated in every state. Three sections of USP <797> deal specifically with ensuring safe CSPs:
Sections of USP <797> also cross-reference other chapters that should be considered when developing practice standards to prepare CSPs:
The pre-administration storage duration and temperature limits specified in USP <797> apply in the absence of direct sterility testing results that justify different limits for specific CSPs.4 CSPs prepared in hospitals are almost always considered medium risk as defined in <797>, since multiple doses are produced from a common source.
In the absence of direct sterility testing, the pre-administration storage times for medium-risk CSPs are 30 hours at controlled room temperature, 9 days at refrigerated temperature, and 45 days in a solid frozen state at -10 to -25 degrees Celsius or colder.5
Extended BUDs for CSPs define the shelf life of the product and in turn can reduce the workload and waste associated with shorter expiration dates. In addition to testing for sterility for extended periods of time, it is critical to recognize that truly valid evidence of stability for predicting beyond-use dating can be obtained only through product-specific experimental studies.4 Semi-quantitative procedures, such as thin-layer chromatography (TLC), may be acceptable for many CSPs. However, quantitative stability-indicating assays, such as high-performance liquid chromatography, would be more appropriate for others. Examples of such CSPs include those with a narrow therapeutic index, for which close monitoring or dose titration is required to ensure therapeutic effectiveness and to avoid toxicity; products where a theoretically established beyond-use dating period is supported by only marginal evidence; or situations when a significant margin of safety cannot be verified for the proposed beyond-use dating period.
Because beyond-use dating periods established from product-specific data based on appropriate instrumental analyses are clearly more reliable than those predicted theoretically, the former approach is strongly recommended in USP <797> in order to support dating periods exceeding 30 days.3,4 When published studies meet all of the components—complete description of materials, test conditions, and methods; use of a stability-indicating assay; performance of a time zero determination of drug concentration; use of replicate assays; and an appropriate conclusion based on results—reviewers can conclude that the results constitute acceptable proof of stability.6 Accordingly, at our hospital we use published stability data to help us determine BUDs.6-9 Some organizations use independent laboratories to perform stability studies, at a cost of approximately $12,000 to $16,000 per CSP, but to avoid that expense, we do not expand beyond a 30-day BUD without having access to published studies.
Process and Batch Validation
When identifying methods to validate product integrity, product sterility also should be considered. Accordingly, we have validated our processes and continue to test each batch of CSPs for sterility.
USP <797> states that technologies, techniques, materials, and procedures other than those described in the Chapter are not prohibited so long as they have been proven to be equivalent or superior with statistical significance to those described therein. To determine statistical significance, we use research articles that identify the standard approach used by pharmaceutical manufacturers to create and validate a process simulation for aseptically filled products.10 Consequently, we determined that 3,000 trypticase soy broth (TSB) samples prepared by our robot should be tested to check for aerobic and fungal contamination.
We also perform fluid thioglycollate (FTG) environmental sampling to test for aerobic and anaerobic growth. As no contamination has been identified, we are confident that our robot process has met a statistically significant burden of proof and GMP standards. If our process changes significantly, the new process will need to be re-validated.
USP <71> defines sterility testing based on statistical methods to detect low-level contamination in a uniformly produced batch; IV robot technology ensures uniform CSP production. We continue to do end product testing for every batch of CSPs. However, based on the results of our process validation study, we are now comfortable releasing CSPs from quarantine three days after drawing the test samples if there are no signs of contamination. Prior to the completion of our process validation, we quarantined all CSPs for 14 days to be sure that there was no contamination in our test samples. USP does not require CSPs to be quarantined.
Each organization must use statistically significant testing to prove that robotically produced CSPs are sterile, because operational processes vary by location. Testing should be performed with media that promotes aerobic, anaerobic, and fungal growth. Changes in process may require re-testing to establish process validation. End product testing must be performed consistently using appropriate sample sizes and products should be quarantined until there is a level of certainty that there are no contaminants in the product samples. Assign extended BUDs conservatively, and only after established stability tests have been documented through the literature or by an independent laboratory using analytical methods and sterility for extended dating has been proven.
If a commercially produced CSP is not available, ensuring that safe CSPs are produced in the pharmacy requires adherence to USP <797> BUD and process batch validation guidance. Continued vigilance is requisite to ensuring an end to compounding errors and noncompliance, which can result in patient harm.