Reducing Error Potential with Carousel Implementation

March 2011 - Vol.8 No. 3 - Page #26
Category: Automated Medication Storage and Retrieval Systems

Attending to continual demands to improve the cost-effectiveness of patient care requires emphasis be placed on decreasing drug inventory and reducing error rates, which makes pharmacy automation a necessity. To realize these goals, we decided to employ carousel technology to provide automated inventory management and additional layers of safety checks. The challenge to such a decision often lies in clearing obstacles that impede the integration of multiple automation systems, not only for smaller and rural hospitals, but for many mid-size and large facilities as well. While proper technology implementation invariably leads to increased efficiency, convincing administration that the costs of implementation will be offset by long-term safety and efficiency gains may not be as straightforward.

To justify the initial cost, it is necessary to demonstrate a positive and realistic return on investment to the governing administration and major stakeholders. At Eastern Maine Medical Center, a 411-bed facility in Bangor, Maine, we considered implementing carousels and pharmacy scanning for many years before the project was approved. When evaluating new automation proposals, approval committees tend to zero in on a single issue—the overall cost of the project. In this case, the total required investment was significant, as the physical plant upgrade costs and ongoing service contract charges had to be included with the hardware price. However, by forecasting reductions in carrying inventory, expired medications, dispensing errors, and medication errors reaching the patient, as well as an increase of inventory turns, we presented a strong case that the project was well worth the upfront cost, along with the yearly expenditure for service contracts and operations.

Replacing an Outdated System
Before implementing the carousel, we used an antiquated ordering system. All of our formulary items were stored on shelves in the pharmacy, and while medications were alphabetized, same-name/different-strength medications were stocked side by side, increasing the possibility that a wrong strength medication could be picked inadvertently. The process for replenishing medications was equally inefficient; our buyer simply walked the rows looking for items in need of replenishment, scanned those items, and input the amount she thought would be needed. If an item moved faster that particular week, she would order extra so that we would not run out. This process was not only time consuming, but its accuracy was variable. Because our purchasing was not based on actual use, but rather on what we thought we would use, it was common to run short of certain items or be left with a surplus of medications sitting on the shelf.

Our prior system included three steps that depended on human accuracy during the dispensing of a medication: the first occurred when the medication was received and placed in its bin, the second when the technician pulled the medication for dispensing, and the third when the pharmacist double-checked the medication. Although this process was managed accurately the majority of the time, any processes that rely on human interaction introduce the potential for error. To mitigate this risk, we focused our efforts on those points in the system that could potentially fail. We looked to carousel hardware and software to improve the dispensing process for technicians, as well as add electronic safety gates that would prevent errors from passing through and potentially reaching patients.

Baseline Data and Error Reports
Our first step was to find out how many and what types of medication errors were occurring. In gathering this baseline data we reviewed two areas in the error reports: the overall number of incident reports per 10,000 doses dispensed for the facility and pharmacy errors that were not caught and were delivered to the floors. The overall incident report rate included not only dispensing errors from the pharmacy but also nursing errors, including wrong patient, wrong dose, wrong route, etc. Also, we looked at baseline data on reported dispensing errors over the three years prior to our implementation. Using our automated dispensing cabinet reporting system and an internal “Good Catch” employee-reporting program (see sidebar), we compiled a report that showed the dispensing error rate for incidents that reached the patient.

Under our previous system, the facility was averaging 44 medication errors per month that reached the patient, an error rate of 2.532 errors per 10,000 transactions. This was below the baseline that was set for our facility of 3.15 per 10,000 doses, but left room for improvement. In addition, we took a three-month baseline of pharmacy dispensing errors. Using this data we found that an average of 34 dispensing errors per month, approximately 256 individual doses, were sent to the floor or refilled in error.

Convincing Staff of the Implementation Value
Attaining staff buy-in for a new system implementation is vital and should be addressed early on in the process. Bringing in new software and changing the landscape of the pharmacy will generate some resistance from staff unless sufficient time is spent in training so that staff members understand the process, what is expected from them, and why change is needed.

In our hospital, the new carousel technology affected all aspects of our technicians’ responsibilities, so we first introduced the software on a test system to instill comfort with the process. We then held training sessions and made resources available for the first month after implementation to address questions as they came up. More in-depth training was given to select members from each shift—those designated as superusers—so that when administrators were not available there would be staff on hand who were able to assist. This fostered a sense of ownership among the technicians and ultimately led to a successful go-live.

We also created a training guide that detailed all the steps of basic transactions. This was included in binders prepared for each individual carousel that also contained the user guide, our downtime manual, our facilities antidote sheet with carousel locations, an inventory listing of medications housed in that carousel with locations, and a sheet with contact information for customer support. These tools have been helpful references, allowing our technicians to manage most problems that arise in the operation of the carousel.

Inventory Reductions
One of the most pressing needs in our facility was finding effective inventory reduction strategies. Our stakeholders were particularly interested in decreasing inventory expenditures, and this was one of the major reasons we looked to carousel automation. We needed to replace our just-in-case ordering process with a just-in-time inventory management system.

We started by reviewing the distributor’s report of purchasing totals for six months prior to our go-live date and calculated a three- to five-day usage total for most formulary items. Package size and the usage report were reviewed to assign minimum and maximum values for each inventory item. By designing the bin locations within the carousel to fit only the maximum amount, we were able to cut down on excessive ordering while still maintaining an optimum amount of stock. We established minimum values so that when a medication reaches the set number a prompt is sent to the buyer to place an order for that item, while enough inventory remains on hand to last until the replenishment order arrives.

About 20% of our inventory—infrequently used formulary items—had to be addressed differently. These items had minimum and maximum values assigned based on package size for tablets/capsules or by quantity for bulk items, such as vials, oral solutions, and topicals. This process took time to develop and organize and was frequently adjusted for about three months following our initial go-live date. We continue to modify and update minimum and maximum values, as well as bin locations for products that change in usage or are added/dropped from our formulary. We have found that tweaking minimums and maximums is a process that needs to be revisited regularly to ensure accuracy.

Reducing Error Potential
The implementation of carousels and scanning software enabled us to add multiple electronic safety checks into the distribution model, and a product scan step was added to our newly implemented dispensing process. After the medication is put into the queue for restocking, it is scanned and put away in the carousel. When the medication has been put in its bin, the bin also is scanned to ensure the drug is in the right location. A third safety check occurs when the medication is scanned before it is dispensed. If the medication is incorrect, the carousel will not let the technician proceed with the order; if correct, the technician initials the medication label and the pharmacist provides the final verification of the medication. The order is then cleared to leave the pharmacy for distribution.

Outcomes: Hard Gains
In the three years prior to our carousel implementation, we were averaging a $1.2 million dollar carrying inventory, with roughly 24 inventory turns per year. After implementation, we were able to cut our inventory to an average of $825,000 in 2010, realizing a 32% reduction. We also saw an increase in the number of inventory turns from 24 to 33 and a reduction in the amount of expired medications in the pharmacy due to improved tracking of expiration dates and increased inventory turns. By looking back on year-end inventory reports for the past three fiscal years and taking an average, we determined an accurate starting point for our data. After going live, we kept monthly reports of inventory totals for the carousels and used these to demonstrate the project’s success to administration.

The implementation of the carousel and the additional safety checks have resulted in a significant decrease in dispensing errors, from approximately 34 dispensing errors per month to 5.5, an 84% reduction. In line with that decrease, misfilled units fell from a preimplementation average of 256 per month to approximately 41, an 84% reduction in per-unit errors. This reduction also positively impacted the overall error rate for our facility. In the first full year after implementation we saw an overall decrease of 80 medication errors that reached the patient as compared with the previous year; this translates to a rate of 1.84 errors per 10,000 transactions, a 32% reduction in medication errors that reached the patient (see Figure 1).

Click here to view a larger version of this Figure

A large percentage of this reduction in error rates was realized from the decline in errors leaving the pharmacy. Every error that is stopped in the pharmacy is one less error that could possibly make it to a patient bin or ADC pocket, and one less incorrect medication that could possibly be administered to a patient.

The Institute of Medicine estimates that at least 1.5 million preventable adverse drug events occur each year in the US.1 Although decreasing costs and lowering inventory are important considerations when making the decision to implement a carousel, the most important consideration is improved patient safety. Automating the process is what made the difference for our hospital in our efforts to decrease dispensing errors to patients.

Outcomes: Soft Gains
Presenting verifiable figures makes it clear to administration that the time and expense dedicated to carousel implementation has produced financial returns, but it is important also to present benefits that cannot be quantified monetarily. One of the first things we noticed after our carousel implementation was how streamlined the pharmacy had become. Technicians no longer had to search for medications on the floors and the increased focus on dispensing resulted in a 60% faster response for our STAT and first-dose orders. Our pharmacy buyer now generates an accurate order report in minutes rather than manually searching for items on shelves, a process that used to take a significant amount of time each day. This new capability has helped to reduce over-ordering, increase efficiency, and streamline the purchasing process. Our yearly inventory is now a quick and seamless process; we can see an accurate report of inventory totals at any time.

Implementation of carousels and other automation systems are an investment in patient safety and staff efficiency. In our hospital, the financial costs of implementation were returned exponentially in both the hard and soft gains the system delivers. As the medication distribution model continues to evolve and incorporate new technologies, it is important to consider not only the immediate financial cost, but also the dividends in improved patient care and safety the system will offer over time. As we continue to revise our distribution model and move toward bedside bar coding, we expect to see error rates reduced even further. Our eventual goal is to have zero preventable medication errors each year.


  1. National Academies Press. Institute of Medicine report. Preventing Medication Errors: Quality Chasm Series. Accessed February 7, 2011.

Michael Hodgkins, CPhT, is the distribution operations supervisor for Eastern Maine Medical Center in Bangor, Maine. He has been system administrator and automation engineer for the pharmacy since implementation of the technology in 2009. Michael can be reached by email at

Encouraging Error Reporting

The Good Catch program is designed to increase the reporting of errors. Pre-printed cards are placed in the pharmacy and nursing units, as well as in other areas where errors may occur. Employees are encouraged to fill out a card when they see an error occur or discover a process that might lead to an error. The cards are filled out with the reporting person’s contact information, the date of the event, the patient’s medical records number, where the event occurred, and a brief description of the event. These cards are collected throughout the week and the findings are sent to multiple committees and departments for review.


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

Current Issue