DrAfter123/DigitalVision Vectors

How AI, VR are improving workforce well-being at Duke Health

Duke Health is using AI and virtual reality tools, developed in collaboration with its staff, to support clinical decision-making and enhance workforce safety.

Workforce well-being has long been a concern of healthcare leaders nationwide. Excessive workloads, clinician shortages and the aftermath of a once-in-a-century pandemic have resulted in high rates of burnout and low levels of career satisfaction among healthcare workers. Healthcare organizations are employing a wide range of strategies to address this issue, including the use of advanced and emerging technologies.

At Duke Health, AI and virtual reality (VR) tools are demonstrating their effectiveness in mitigating factors that negatively impact workforce well-being. While AI is being used to support clinical decision-making and staffing operations, VR is helping train clinical staff to manage potentially violent interactions with patients, Duke Health's Theresa McDonnell, DNP, shared with Healthtech Analytics.

McDonnell, senior vice president, chief nurse executive and vice dean for clinical affairs and innovation at Duke, also emphasized that the tools were developed in close collaboration with clinicians. This not only helped the health system identify the most pressing workforce well-being challenges that needed to be addressed, but also made gaining clinician buy-in easier.

Using AI to support clinical workflows

Several factors impact workforce well-being, causing stress and anxiety among healthcare workers. One of these factors is the high demands of healthcare that often clash with dwindling resources, creating environments where clinicians may make mistakes.

"We've commonly referred to this syndrome as burnout, but when we dig into it, it's really that constant feeling of fear of not being able to deliver what a patient needs, fear of making a mistake, the anxiety of constant interruption and the things that really detract humans from concentration and work," said McDonnell.

Research shows that when clinicians are anxious and fatigued, their clinical decision-making suffers. According to a study published in 2025, a machine learning model found that signs of fatigue in clinical notes correlated with worse physician decision-making

To address workplace conditions that drive anxiety and fatigue and increase the odds of lower-quality patient care, Duke Health has turned to AI.

McDonnell shared that the health system is utilizing a patient deterioration index tool that leverages AI algorithms to alert clinical teams when a patient's health suddenly takes a turn for the worse.

"[The tool] notifies a rapid response team that can then check in with the primary care team and bring resources long before we have an urgent or emergent situation unfolding with the patient, leading to better patient outcomes, while also supporting the frontline team much sooner and helping avoid those emergency high-intensity situations," she said.

The health system is also using AI to mitigate conditions that commonly result in rapid deterioration, like sepsis. An early sepsis detection tool allows clinical teams to prevent or head sepsis off in its early stages, McDonnell said.

There are also several new AI tools Duke Health plans to implement. For instance, the health system is working with third-party vendors to implement AI-powered patient rooms. McDonnell shared that the system is refreshing three units at Duke University Hospital, implementing fiber optics and computer vision technologies to enable early detection of patient falls. The technologies would be able to detect a potential patient fall and alert clinical teams to intervene.

In addition to these rooms, Duke Health is also partnering with an external vendor to roll out an AI-driven system for workforce management.

"Rather than a charge nurse sitting there and trying to block and tackle or a staffing committee, we'll have advanced analytics to help us match staff needs and staff preference against the needs of the unit," McDonnell said.

Using VR to curb workplace violence

Workplace violence is rife in healthcare, with Bureau of Labor Statistics data showing that healthcare and social assistance experienced the highest rates of workplace violence of any private industry sector in 2021 and 2022.

"Workplace violence is not new," McDonnell said. "We have always cared for patients who have medications on board that can alter their mental status. We have always had patients who have conditions like dementia or Alzheimer's, that can create risk. What we lack, though, or what we have lacked and what we have seen decline, is the preparation of the workforce."

At Duke Health, VR tools are helping prepare the workforce to manage and de-escalate situations with the potential to turn violent.

McDonnell explained that the health system has supported the development and utilization of VR goggles that can train staff through various simulated scenarios. In real time, staff can practice their responses to an escalating scenario in a safe VR setting.

"That, too, helps with well-being -- preparation," she said. "If you don't have the fear of anxiety of not knowing how to be prepared or react to a situation, that lends itself to better well-being."

Keeping clinicians at the forefront of tool development, deployment

Clinician involvement was critical to the success of AI and VR tools at Duke Health.

"Anyone who knows about integrated facilities design knows that it really succeeds when you have full engagement and input from the people who are closest to the work," McDonnell said.

The health system included clinicians in tool development and deployment in various ways, including through hack-a-thons. During these sessions, staff were encouraged to come forward with ideas to solve specific problems.

Not only that, but Duke Health also launched a campaign called 'Improvement through Innovation,' where staff were provided with a QR code that they could scan to access a short survey. The survey had four questions: What is the problem? What do you see as being the solution? What's been tried? How can we help you?

"Leaders always want to come to the table with solutions," McDonnell said. "Somebody comes to me, I want to fix it for them. It's what we want to do. We're humans. But the best thing any leader can do is ask a lot of questions and help empower the people closest to the work to solve the problems."

For instance, when designing the new rooms with patient fall detection technology, the health system included frontline physicians, nurses and support staff in all design sessions with the IT teams and external vendors.

As Duke Health moves ahead with its AI and VR integration plans, leaders are keeping tabs on the impact of these tools. McDonnell said that the health system is measuring various key performance indicators, including staff overtime and employee retention.

They also regularly conduct an internal survey called Culture Pulse. According to McDonnell, the most recent survey indicated increases in staff members' sense of belonging and satisfaction.

McDonnell hopes to share Duke's workforce well-being efforts with other health systems, emphasizing not only the potential of advanced technology to improve well-being but also the importance of including staff in decisions about how and where advanced tools are used.

"[We present] unique opportunities for staff to really be involved in the integrated design of solutions that hopefully will not only transform care here at Duke, but we can then share those learnings with others," McDonnell said.

Anuja Vaidya has covered the healthcare industry since 2012. She currently covers the virtual healthcare landscape, including telehealth, remote patient monitoring and digital therapeutics.

Dig Deeper on Artificial intelligence in healthcare