New & Improved!
Register for our 6/21 Webinar: USP Compliance: USP <795> Facility Requirements

Optimizing Central Pharmacy Dispensing Technology
May 2016 - Vol. 13 No. 5 - Page #2

The pros and cons of centralized versus decentralized dispensing models have been debated for many years in hospital pharmacy. Pharmacy leaders have created return on investment models for automation and other technologies to support both models, and robust debate has occurred in the pharmacy community regarding the safest and most efficient combination of process, human resources, and automation to strategically position hospital pharmacies for success in a value-driven model. To this end, the University of Rochester Medical Center (URMC) pharmacy developed a strategic plan, which supports patient-centered clinical activities that enhance value-added services for the hospital (see FIGURE 1), to safely and efficiently move medications from the dockside to the bedside.

However, developing a strategic plan is only the first step; optimizing the medication distribution model and technologies employed is required. Various distribution strategies, including implementing robotic drug storage, automated dispensing cabinets (ADCs), bar code verification, and improving human workflow systems must be considered. Moreover, the efficiency of such strategies, as well as the smart implementation and maintenance of technologies, also are critical considerations.

Developing Effective Centralized Drug Distribution

Hospital operations require a complex integration of multistep procedures; thus, many hospitals utilize some degree of a hybrid drug distribution model, and URMC is no exception. Within the URMC health system, most hospitals are 90% decentralized, with medication distributed on the units from ADCs, but our largest facility, Strong Memorial Hospital (SMH), sends approximately 40% of medications to the unit through ADCs. Therefore, about 60% of SMH’s medications are dispensed using a centralized distribution model. Effective management of these medications is essential, as failure to monitor and maintain optimal efficiency in the central pharmacy can quickly invite disorder.

Improving Safety

At the crux of every technology discussion is patient safety—specifically, how can technology be leveraged without negatively impacting safety, or more importantly, in order to improve patient safety? Although initial safety projections from vendors that provide centralized pharmacy dispensing technology may be compelling, post-implementation metrics are required to substantiate a patient safety argument.

It is important to remember that while the benefits of these technologies provided support for the acquisition decision, ongoing measurements of success are equally important to justify our financial strategy and reinforce the safety benefits. Many of the predictive metrics included financial goals related to increasing inventory turns, reducing waste, and increasing efficiency. Measuring safety metrics are more challenging, but crucial.

For example, when implementing bar code medication administration (BCMA) systems, literature is available that supports error reduction, but metrics also must be measured at our own hospital sites. A study conducted at SMH to evaluate medication errors following implementation of an integrated electronic medical record (EMR) and BCMA system reported an initial 25% reduction in drug administration errors in the first year, followed by a 50% reduction in the second year (see FIGURE 2).1 Ongoing studies suggest even greater reductions, with success significantly tied to our >97% reported compliance with BCMA, which highlights the importance of utilizing the technology to its fullest extent.

A prerequisite to BCMA success is ensuring that 100% of medications leaving the pharmacy will scan correctly. Centralized pharmacy technology can play a role in driving success by ensuring that bar codes have been applied to all products. In addition, tracking metrics related to safety and efficient use of the technology in centralized distribution models is critical.

Considerations for Choosing Central Pharmacy Carousels

To support the URMC strategic plan, in 2008 the pharmacy implemented carousels in the central pharmacy and upgraded its dispensing robot. When choosing centralized automated storage and retrieval devices for the URMC pharmacy, decision support included three primary safety considerations: single product accessibility during removal; forced bar code scanning, including a solution to add bar codes to items if needed; and expiration date tracking. Efficiency concerns included the speed of the device, the ability to queue several transactions, the use of inventory management software to adjust stock levels, and automated ordering creation. We also considered logistical aspects of the automation, such as whether the device was modular and expandable, whether it could be used for refrigerated products, and the ability to link inventory management across multiple facilities (see TABLE 1).

These devices are utilized at URMC to automate the ADC refill process through an interface that communicates pocket quantities to the central stock devices, with the goal of reducing or eliminating human error in this process. Medications are then automatically presented to the user in a single storage container based on refill needs, scanned upon removal, and subsequently scanned at the ADC device for confirmation upon restocking.

Upgrading the Centralized Robotic Dispensing System

Concurrent with the evaluation of automated carousel storage and retrieval systems, the URMC pharmacy also upgraded its central robotic dispensing system. At the time, most bulk item repackaging into robot-ready, unit-dose packages was done onsite with little automation. This process was exceedingly labor intensive, requiring multiple checks by technicians and pharmacists and significant time to load these medications into the robot. Thus, the efficiency of cart fill processes had dropped to below 60% of unit-dose products being picked by the robot, with no significant first doses being filled automatically by the system (<10%); frequent downtime increased these inefficiencies. Our initial goal was to increase unit-dose fill rates to over 80%, an aggressive objective considering that the health system includes a large pediatric hospital as part of the dispensing model and incorporates first-dose fills. In addition to increasing efficiency, improving the safety of the repackaging process was a primary goal.

The upgraded robotic system, implemented in 2008, offered many advantages to meet our safety and efficiency goals, including bar code verification, safeguards and failsafe processes, pharmacist verification, flexibility, redundancies, return efficiencies, and quality assurance (see TABLE 2). Given these safety improvements, the metrics that were most critical for us to track included error rates and percent fill rates. Tracking error rates for wrong product packages selected by the system involved checking 100% of the products dispensed, followed by quality assurance checks as needed or when indicated due to system upgrades or downtime. Three months after implementation, while adding additional patient beds over the interim, over 200,000 doses dispensed by the robot were manually checked without detecting any errors. By early 2016, the system had selected over 16,000,000 doses without evidence of any package selection errors; the only errors detected in this time frame were rare packaging anomalies, such as a tablet fragment being packaged instead of the full tablet.

Given the accuracy of the robot medication selection and packaging process, much attention has been focused on increasing fill rates and incorporating as many products into the automated dispensing process as possible. Our strategy includes balancing the most efficient use of ADCs, including focusing decentralized storage on less predictable use products, typically PRN medications, to reduce return rates into the robot and then using this production capacity for more predictable scheduled medication doses.

Click here to see TABLE 2.


ADC refill errors data, accessed 2 years after optimizing our central pharmacy dispensing technology, demonstrated a 50% drop in refill errors overall and in the number of ADC refill errors reaching the patient. Five years after maximizing the automation (with BCMA implemented in the interim) we experienced a 75% drop in reported ADC refill errors overall, with no reported ADC refill errors reaching the patient (see online-only FIGURE 3). In addition to this direct measurement of error reduction, we were able to redeploy staff to other critical functions in inventory management (eg, stock rotation, expiration date tracking, ADC stock level adjustments, additional checking processes for inventory not included in automated solutions, etc), given the approximate 40% reduction in time required for tasks such as moving inventory into stock, removing medications for ADC replenishment, and medication reordering.

With the increased capacity, efficiency, and safety of our systems, we were able to easily meet our goal of >80% central pharmacy unit-dose automated dispensing and significant first-dose dispensing. Current automated fill rates are over 85% of unit-dose products and approximately 70% of all first doses (see online-only FIGURE 4). This success has been achieved over a 7-year period where we have measured steady growth, with an over 43% increase in doses dispensed. Throughout this period, we have continued to minimize the time required for pharmacists and technicians to engage in manual medication distribution activities, as maintaining our efficiency goals is critical to our overall strategy.


Final Thoughts

The goal of this efficiency evaluation was not only to demonstrate that the technology-assisted model is safer (although we have certainly accomplished that), but also to provide the data necessary to optimize this model. The current focus is now on leveraging this technology to create the safest and most efficient model for the medication distribution system. This includes supporting important patient safety initiatives at the bedside, such as BCMA, by maximizing compliance in scanning medications as a safety check prior to administration. In turn, this strategy facilitates more focused time on new and improved value-added services that pharmacy staff can spend in patient-care pursuits, and enables the medical teams to improve outcomes in patient care and help develop our future clinicians. A drop in automation efficiency would pull already limited resources away from these roles, so remaining diligent in tracking the important metrics that drive these efficiencies will continue to be the foundation of our future success.


  1. Adams MJ, Reagan P, Haas C, et al. Electronic Medical Record Implementation Associated with Immediate Decrease in Self Reported Medication Errors: A Comparison of Similar Time Periods Pre– and Post– Implementation. Proceedings from the combined American College of Preventive Medicine/American College of Medical Quality meeting. February 23, 2012: Orlando, Florida.
  2. Cina JL, Gandhi TJ, Churchill W, et al. How many hospital pharmacy medication dispensing errors go undetected? Jt Comm J Qual Patient Saf. 2006;32(2):73-80.

David F. Webster, BS Pharm, MSBA, is the associate director of pharmacy operations and director of the PGY1 pharmacy residency program at University of Rochester Medical Center, Rochester, New York. He received a BS in biochemistry from Saint Bonaventure University, a BS in pharmacy from the State University of New York at Buffalo, and an MSBA in operations management from the Simon School of Business at the University of Rochester.


Potential Consequences of Human Error in the ADC Fill Process

Peer-reviewed data validates URMC’s internal observational data regarding human error in the fill process. One study reported a hospital pharmacy technician fill error rate of 3.6%, but also a 79% pharmacist-check interception rate of these errors, thus allowing approximately 0.75% of doses dispensed in error to leave the pharmacy undetected.2 Although a 99.25% accuracy rate may be acceptable in some industries, extrapolating this error rate to our automated dispensing data, a manual system for these unit-dose products would have resulted in over 500,000 errors, with over 120,000 errors leaving the pharmacy undetected in this 8-year period. Clearly, robotic medication dispensing offers significant safety benefits over manual processes. Moreover, all URMC bulk tablet/capsule repackaging to unit-dose products is completed using this system, which ensures that all unit-dose packages have a readable bar code for bedside scanning—even products that are not stored and dispensed by the robot (eg, controlled substances).

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:

    As of May 24, 2018,'s Terms of Use and Privacy Policy have been updated.
    You can review the Terms of Use HERE and the Privacy Policy HERE.
    By continuing to use, you are agreeing to these changes.