Success at Guthrie Clinic Leads to AI Testing and Expansion
At Guthrie Clinic, which serves parts of rural New York and Pennsylvania, an AI-supported telesitting program was generating measurable results by spring 2024. The clinic’s use of the Artisight platform takes data from the cameras and microphones set up in patient rooms to analyze and augment clinical activities in real time.
Many protocols that require another clinician to sign off previously took 30 minutes to complete. Now, the process takes 60 to 90 seconds via camera, and the nurse never needs to leave the patient’s side. Nurse satisfaction has risen, and turnover has fallen from a high of 25% to 13%.
“Just being able to recruit nurses and maintain their employment is an indicator that things are working,” says Terri Couts, senior vice president and chief digital officer at Guthrie Clinic. “I’ve seen a lot more teamwork and integrated communication.”
Couts, who is also a registered nurse, and her team plan to further leverage AI to enable better patient care.
“Eventually, as we continue to employ the ambient listening feature, we may be able to get rid of the keyboard in the patient room, enabling clinicians to be more present,” she says.
The team is also testing the use of AI to see, hear and record specific events in a patient’s medical record. For example, immobile patients can develop pressure ulcers, or bedsores, if they are not turned every two hours. The ICU nursing team is training the AI with verbal cues and deliberate movements that are then documented in the electronic health record. The goal is to train the computer vision to see this procedure and eventually capture and record it without the verbal cue.
DIVE DEEPER: Why is Patient Room ‘Next’ the evolution of care delivery?
OhioHealth Standardizes Patient Discharging with AI
OhioHealth implemented AI to improve and standardize the discharge process at its two largest hospitals. Scott Estep, system vice president of nursing operations and capacity management, led the initiative.
Estep is a trained nurse, but also has a background in process improvement, which serves him well in managing the operations side of healthcare. “I’m a data guy. I like processes,” he says.
During an analysis of the Central Ohio-based health system’s discharge process, Estep says they realized that each member of the clinical multidisciplinary team had varying ideas on the patient discharge process. “When asking frontline caregivers when they thought their patient might go home, you’d get different answers depending on who you asked, even when looking at similar patients,” he adds.
OhioHealth collaborated with a healthcare automation platform company on an AI-powered custom solution that focuses on improving the discharge process by elevating the elements of care that pose barriers to discharging a patient to the right members of the care team.
“We spent nine months co-designing this tool,” Estep says. “Our entire multidisciplinary team collaborated closely with the vendor throughout the entire development of the product, which ensured we had something that met the needs of our team and patients.”
READ MORE: Demystify artificial intelligence adoption for your healthcare organization.
The tool is now at work in two hospitals, used mainly by the clinicians most involved in patients’ hospital stays. The platform analyzes patient data upon check-in and adds various input from staff along the patient journey. It then generates an estimated date of discharge (EDD). Estep calls this a “shot clock,” a proposed check-out timer that starts counting down. Staff members can change the EDD if they disagree, but then challenge themselves to beat the date.
“To date, we’ve had a 17% reduction in excess hospital days,” Estep says. “If we reduce excess days, we shorten the length of stay for our patients. That also creates much-needed bed capacity. From this work, each day, we create an additional 23 beds that we didn’t have before.”
Having a shared discharge goal and visibility to information helps nurses, providers and other staff members feel empowered in their patients’ care. The next step is to grow the program. “Right now, we’re only at two hospitals. After six to nine months of data, we’ll decide where to expand to next,” Estep says.
