Omaha-based Nebraska Medicine is deploying artificial intelligence to predict length of stay and streamline discharge workflows, which is freeing up inpatient beds and easing administrative workload for nurses.
Nurses at Nebraska Medical Center are using AI to support discharge planning, Kelly Vaughn, MSN, RN, the system’s chief nursing officer, told Becker’s. Among its uses, AI helps identify patients ready for discharge who are waiting on final steps and may be candidates for placement in the hospital’s departure lounge.
Historically, nurses manually reviewed charts to determine which patients met the criteria to be sent to the lounge. Now, AI supports that step by reviewing the chart more frequently and efficiently to flag patients who may meet established criteria. Nurses then review flagged patients for inclusion or exclusion, with the ability to see exactly where the AI surfaced supporting information in the chart.
“That’s been highly, highly effective,” Ms. Vaughn said.
Reducing the need to search through documentation manually has helped free up beds faster, she said. In December alone, the departure lounge saved the hospital an estimated 135 inpatient days.
AI is also being used to help predict length of stay based on a patient’s medical diagnosis and what physicians have noted in the record. Ms. Vaughn said this information has allowed clinical teams to begin discharge planning conversations earlier with patients and families. Length of stay reductions have created capacity equivalent to 50 additional inpatient beds per month.
“It’s been incredible to create more and more access for patients,” she said. “Those beds have been immediately filled, so it’s been really good for our patients and staff to have that availability.”
A key factor in nurse adoption and trust of AI has been Nebraska Medicine’s collaborative, nurse-led development process, Ms. Vaughn said. The health system pairs nurses directly with forward deployed engineers — technical staff who observe front-line workflows, ask questions, identify patterns and build AI-based tools designed to reduce friction in daily tasks.
“Typically, what you’ll see is the opposite,” Ms. Vaughn said. “A solution is created and it’s inserted into a nursing workflow, and the nurse adapts their flow to whatever that system has created.”
So far, clinical staff and engineers have co-developed about 50 use cases.
“[Nurses have said] this isn’t like other systems where I’ve had to adopt it, where it’s been burdensome and can create a heavier workload,” she said.
