Tech Explained: Here’s a simplified explanation of the latest technology update around Tech Explained: Using AI to balance nursing workloads in infusion centers in Simple Termsand what it means for users..
The nursing workforce continues to face high levels of burnout, leaving the U.S. health system facing a projected shortfall of 63,720 full-time nurses in 2030. Health systems are working to ease nursing burdens in various ways, including through emerging technology.
At the University of California, San Francisco (UCSF) Health, AI-based technology is helping enhance nursing workflows and experience in its infusion centers. The health system is leveraging an AI-driven capability to help charge nurses distribute the workload more evenly across staff.
“It wasn’t just about making sure there was a balanced [patient] throughput through the infusion center; it was really , how do you take those patients and make sure the nurses taking care of those patients have balanced assignments,” said Marisa Quinn, director of nursing for infusion services at UCSF Health, in an interview.
The AI capability is helping reduce workload inequities among nurses, she added, resulting in lower burnout and higher nurse satisfaction. However, to achieve these outcomes, leaders had to secure and maintain nurse buy-in and trust in the technology.
Understanding the AI-based tool
AI is not new to infusion services staff at UCSF Health. The health system has been using LeanTaas’ iQueue for Infusion Centers for the past decade, which is an AI-based solution.
According to Quinn, the solution helps the infusion services department schedule and level-load patients to prevent bottlenecks in patient flow. Capacity challenges are common in infusion centers. Whilecenters are generally open eight to 12 hours a day, certain times of day may be overwhelmed with patients.
“If you talk to any infusion operator, they’ll tell you that the times between 10:00 and 2:00 or 10:00 and 3:00 are going to be your busiest, highest peak times because patients are coming out of their doctor’s visits and they want to flow right into their infusion care,” Quinn explained. “They don’t want to drive late at night. They want to be home before the traffic. So, all of these variables play into a system that creates high peaks in the middle of the day.”
The AI solution allows UCSF Health’s infusion services team to create schedules that help prevent those peaks and instead spread patients out throughout the day. The system analyzes demand trends and capacity data and suggests schedules to ease patient throughput challenges.
“[The] result is much better throughput and the ability to respond to changes during the day,” Quinn said.
However, as the outpatient oncology environment evolved and the patient population became more acute, just focusing on scheduling and level-loading wasn’t enough. Quinn began hearing from staff that the patient assignment process wasn’t quite balanced, and as a result, some nurses had a heavier workload than others.
This is where the patient assignment capability comes in. UCSF Health began using this capability within the iQueue solution about two years ago.
The AI-driven capability allows nursing leaders to be more intentional about patient assignments, Quinn said. The tool ingests data from the workforce management system, including nurses’ employee IDs, names, shifts and the units they’re assigned to.
The user can ask the system to assign all the patients to nurses for the day or ask who the next patient should be assigned to, she continued. Analyzing workforce data alongside patient throughput and capacity data enables the solution to provide an informed patient assignment suggestion that the charge nurse can either accept or reject.
How it mitigates nursing burdens
The primary benefit of the AI-driven patient assignment capability is ensuring that the nursing workload is balanced throughout the day and across the team, Quinn shared.
The UCSF infusion services department identified these benefits during a pilot before deploying it at scale. Department leaders pulled metrics from the tool and conducted pre- and post-implementation surveys with the nurses. The analysis showed improvements in workload balance and productivity. Not only that, but there was an improvement in nurses’ perception of patient assignments.
“So, 75% said they had better pacing of their assignments after, and that was really what we were looking for,” Quinn said. “This wasn’t a cost savings measure. This was really a question of how to create a more sustainable workforce. How do we keep mitigating that burnout that comes with that fast-paced, intense workload?”
The AI-driven patient assignment suggestions are largely implemented, though the charge nurse is free to defer them if they have information that the system doesn’t. For example, if a patient has an unpredictable reaction to their infusion and requires more care, charge nurses can reject the solution’s suggestions and make adjustments as needed, Quinn said. The solution provides real-time visibility into nurses’ workflows, making it easier to reassign nurses to help their peers and even out the workload.
Deploying the tool with buy-in
Staff buy-in is critical to the success of technology deployments. For AI-driven patient assignment, though, it was about solidifying trust in the technology.
Quinn shared that when the problem of unequal workloads was first identified, staff wanted a technology solution to display acuity levels.
“I didn’t find that there were any evidence-based tools out there — oftentimes it’s the easiest patient that you think is coming in who ends up being the most workload effort for the nurse,” she said. “So, they asked for an acuity tool, and I said, ‘Well, we can look into that, but we have this opportunity to pretty much develop this patient assignment tool with LeanTaaS, so why don’t we start there?'”
Infusion services leaders invested heavily in building trust around the new capability and its ability to ease workload burdens. Quin noted that her team focused on communication and education around how the capability was solving the problem.
Not only that, but leaders also emphasized that the new capability keeps the human in the loop, with the charge nurse being the final decision-maker regarding nurse workload.
“Their clinical judgment was still going to be very much first and foremost, with the system to assist them and not automate their workflows,” Quinn said. “I think anytime you come in and talk about changing nurse workflows, you’ve got to be really careful about how you do that and make sure that they can understand the potential benefit and have feedback in the process if that’s not being realized.”
To that end, Quinn and her team held weekly meetings with staff during the deployment to gather their feedback on the capability and potential improvements.
Quin emphasized that, ultimately, supporting nurse staff and preventing burnout is about listening to them and working with them to identify solutions to their problems.
“We are really passionate about making sure that we’re preventing burnout with everything that we can do in our power,” Quinn said. “And that’s acknowledging the challenges that we are experiencing nationwide across health systems and across nursing, but also locally, too. And so, what can we do at the local level to make sure that our staff feels supported? And feeling supported really comes down to being heard. And I mean, it’s not rocket science.”
Anuja Vaidya has covered the healthcare industry since 2012. She currently covers the virtual healthcare landscape, including telehealth, remote patient monitoring and digital therapeutics.