Skip to main content

Pharmaceutical Robotics Require Careful Assessment

By Greg Freeman  
   April 16, 2012

This article appears in the April 2012 issue of HealthLeaders magazine.

Pharmaceutical robotics has made its way from futuristic, cutting-edge technology to almost becoming just another highly sophisticated tool that we expect to see in larger healthcare facilities. But that does not mean it is a perfect solution to medication dispensing problems. Before you sign on to a multi-million-dollar pharmaceutical robotics program, consider the potential pitfalls you may encounter on the way to all the benefits.

The benefits can be significant. Leaders at the University of California, San Francisco Medical Center, say their robotic system has greatly improved safety. With a new automated hospital pharmacy—believed to be the nation's most comprehensive—UCSF is using robotic technology and electronics to prepare and track medications with the goal of improving patient safety.

Pharmacy processes are labor intensive, explains Sheila Antrum, RN, MSHA, UCSF Medical Center's chief nursing officer. The pharmacy dispenses about 10,000 doses a day of 3,000 medications, and there are about 10 steps to be completed between the order being written and the patient getting the medication, she says. Each step in that process includes a risk for error, and automating the system has taken that risk to nearly zero. Not a single error has occurred in almost 3 million doses of medication prepared since the hospital installed the PillPick system from Swisslog, a global supplier of integrated logistics solutions with U.S. offices in Denver.

UCSF opened a centralized, robotic pharmaceutical production facility in the Mission Bay area of San Francisco in October 2010. Everything is barcoded, which is the first step in eliminating human error, says Lynn Paulsen, PharmD, director of pharmacy practice standards for the University of California Office of the President. Housed in a tightly secured, sterile environment, the automated system prepares oral and injectable medicines, including toxic chemotherapy drugs. In addition to providing a safer environment for pharmacy employees, the automation also frees UCSF pharmacists and nurses to focus more of their expertise on direct patient care, taking them out of the mechanical steps of dispensing drugs, which Paulsen notes is not the best use of their intellect.

"The automated pharmacy streamlines medication delivery from prescription to patient," Paulsen says. "It was important to develop a system that is integrated from end to end. Each step in safe, effective medication therapy—from determining the most appropriate drug for an individual patient to administering it—is contingent on the other."

The new pharmacy currently serves UCSF hospitals at Parnassus (568 licensed beds) and Mount Zion (91 licensed beds) and has the capacity to dispense medications for the new UCSF Medical Center at Mission Bay (289 licensed beds), scheduled to open in early 2015. As the phase-in continues, additional steps in the process will be eliminated as doctors begin inputting prescriptions directly into computers this year.

Once computers at the new pharmacy electronically receive medication orders from UCSF physicians and pharmacists, the robotic system will pick, package, and dispense individual doses of pills. Machines assemble doses onto a bar-coded thin plastic ring that contains all of a patient's medications for a 12-hour period. Nurses use bar-code readers to scan the medication at the bedside, verifying it is the correct dosage for the patient.

The automated system also compounds sterile preparations of chemotherapy and nonchemotherapy doses and fills IV syringes or bags with the medications. An automated inventory management system keeps track of all the products, and one refrigerated and two nonrefrigerated automated pharmacy warehouses provide storage and retrieval of medications and supplies.

Antrum cautions that hospital leaders should assess how the robotic system affects the entire process, not just the mechanical filling of the prescription. That means gauging how the use of robotics affects the delivery of medications, accuracy, and the ultimate effect on patient care, she says. She also recommends getting the end user involved in the development process.

"The main thing we've learned is that you have to figure in the culture of your workplace," Antrum says. "You have to start looking at your processes first and determining how to change them with the addition of this robotic system, rather than introducing the system and then thinking people will adapt their processes. It is vitally important to consider the workarounds that people have built into the system in order to get things done."

Training will be necessary, and Antrum cautions that the manufacturer's training may not be sufficient. The number of people you train also will be an important decision. Some providers may elect to train only certain pharmacy technicians and authorize them to use the robotics, and others will open the training to all pharmacy staff.

"But the more people you elect to train, the more important it is that you ensure uniformity in your training. The human factor never goes away, even with a system like this, and that human factor is what can cause things to go wrong," Antrum says. "People like to make their own workarounds or do things slightly different, but that kind of deviation can be the real weakness in your system. Your training has to address that."

Cost benefits can change over time
One potential pitfall with pharmaceutical robotics is that the benefits of the system may be diminished if the hospital changes the medications it commonly uses. That is the situation at Memorial Health System in Colorado Springs, CO—which has three hospitals and 650-plus beds. Memorial  acquired the IntelliFill i.v. robot from FHT in Daytona Beach, FL, about five years ago, and that led to reduced labor costs in the pharmacy while increasing accuracy and dose tracking, and improving patient safety with less touch contamination.

Over the past five years, the i.v. robot has saved the health system at least $1 million, says James Lewis, PharmD, BCPS, director of pharmacy with Memorial Health. But now that financial picture is changing. The robotic system was attractive at first because it could reduce the expense associated with one particular anti-nausea drug that was quite costly. With the robot compounding the drug, smaller doses could be used and frozen. The savings on that one drug essentially paid for the robotic system, Lewis says.

That product has now since gone generic, so there is no cost benefit from using the robot anymore. The robot is still used to a lesser degree for compounding, and it offers benefits other than cost savings, but the ongoing expense of using it is becoming a problem.

Fast delivery
Filling prescriptions isn't the only job for pharmacy robots. At the University of Maryland Medical Center in Baltimore, pharmaceutical robots travel the hallways and even use the elevators all by themselves as they deliver medications to three critical-care hospitals. When they arrive at their destinations, they announce themselves and then allow authorized personnel access to the medications they are delivering.

UMMC, which has 757 licensed beds, began using the TUG robots from Pittsburgh-based Aethon in 2002, first with just one automated medication cart that delivered orders to the hospital's shock trauma unit from the satellite pharmacy, explains Marc Summerfield, RPh, MS, director of pharmacy at UMMC. The medical center now has eight TUG units delivering orders filled by the pharmacy to the three critical-care hospitals within UMMC.

Hospital leaders have been pleased with the success of the system. The robots have reduced by 60% the time it takes between the pharmacy's receipt of a drug order and its delivery to the nursing unit. Delivery to the shock trauma units has been cut from about 45 minutes to about 25 minutes. Previously, pharmacy technicians had to carry the filled orders to the units, which inevitably took longer than a single-focused robot dedicated to its task.

"There was always some variability with people, but the TUGs just go and come back," Summerfield says.

There have been learning experiences along the way. When first adopted, the TUGs used a different pharmacy software that required nurses to take responsibility for drugs when they arrived at the nursing unit. Additionally, controlled substances could not be delivered by the TUGs and still had to be hand-carried to the units. Both of those problems decreased the usefulness of the robots, but those problems were solved when the TUGs were switched to MedEx, a pharmacy-specific software system that provides a record of the chain-of-custody of all drugs, including controlled substances, explains Summerfield.

Another unexpected problem involved one thing that humans can do but the TUGs can't: Step up an inch or two. That becomes a problem when the hospital's elevators don't stop completely level with the floor, a not-uncommon occurrence in an older building.

"An older infrastructure is something you don't usually consider, but you should," he says. "You're introducing the most modern piece of moving machinery and it expects its environment to be pristine. If your hospital is older, the floors aren't perfect and the elevators aren't always precise. It can be a problem you'll have to overcome."

Drug orders are tagged with a radio frequency identification label and placed in drawers on the TUG, which is then sent on its way. The robot travels the hallways, summons the elevator when needed by sending a radio signal, and announces itself when it arrives on the unit. Then a nurse places a thumb on the thumbprint reader to verify identification.

When the TUG is satisfied that the right person is ready to receive the drug delivery, it opens the drawer containing that unit's medications. The nurse removes the drugs labeled for that unit and places them in a lockbox.

Nurse satisfaction with the pharmacy improved by 23%, delivery reliability improved by 23%, and delivery predictability went up by 50%.

"When I see a TUG going down the hallway, I think of what would be happening if we didn't have the robot," Summerfield says. "Either that med would still be sitting in the pharmacy waiting to be delivered, or a technician would be delivering it and not be at the pharmacy working on other orders."

Greg Freeman is a contributing writer for HealthLeaders Media.


This article appears in the April 2012 issue of HealthLeaders magazine.

Reprint HLR0412-6



Tagged Under:

Get the latest on healthcare leadership in your inbox.