Virginia-based nonprofit health system Sentara announced last week that it has started using artificial intelligence technology to assist with patient monitoring and  discharges. 

Regard is an AI system that integrates with electronic health records. Co-founder and president Nate Wilson said it’s used by more than 150 hospitals across the country. Dr. Rollin Reeder, chief health information officer at Sentara, said that all 12 of the health system’s hospitals now use Regard. 

The tool helps physicians and case managers more easily parse through patients’ hospital admission charts, monitoring notes during their stay and their medical histories, both men said. The goal is to help diagnose new or worsening symptoms sooner and think proactively about patients’ discharge times and follow-up care. 

“If a patient is going to need rehab or help at home, you want that on the table at admission, not the afternoon they’re medically cleared to leave,” Reeder said. “Catching it early gives the case manager time to find a bed, arrange services and walk the family through it.”

Discharge planning begins as soon as a person is admitted for inpatient care. Case managers analyze health data about the patient to help determine when it may be safe for them to go home or whether they may need to transfer to a rehabilitation facility first. 

Case managers also begin the work of determining what services are covered by patients’ insurance. But the information is scattered, Reeder explained. 

“Most of those answers are buried” in patients’ charts, medical history, screenings and doctor’s notes, he said. 

“The hard part has always been that it’s spread out everywhere, so case managers spend a lot of their day clicking around just to get the basics before they could start planning,” Reeder explained. 

Regard helps scan and filter information, Reeder said, and case managers appreciate a feature that allows them to pull up rundowns of what may have changed with the patient since the previous day. 

Sentara emphasizes AI is a tool, not human replacement

A press release announcing the adoption of Regard implied that it’s estimated to save “nearly 90,000 hours of staff time – the equivalent of hiring 40+ full-time employees.”  

When asked if use of the technology may lead to reduced case manager staff, Reeder said “no one is losing their job over this.” 

Staffing levels and certain services have been in flux for some hospitals and clinics over the past year following H.R. 1, a reconciliation bill Congress passed last summer that entails changes to hospital funding mechanisms and puts roughly 300,000 Virginians at risk of losing Medicaid insurance. 

Citing the bill as a contributing factor, Valley Health changed staffing contracts and trimmed services this past spring. Last winter, Centra closed its labor and delivery unit at a hospital in Farmville. Last fall, Shenandoah Valley’s Augusta Health closed three clinics. 

Virginia hospitals amplify risks federal shifts pose to patient care

Thirteen rural hospitals — including a Sentara hospital in Halifax —  have also been identified as vulnerable to service cuts or closure over time, a state report indicated. 

Reeder said the estimate of 90,000 hours Regard saves being equivalent to hiring more than 40 full-time staff was meant to reflect how case managers can spend less time glued to a computer screen and more interfacing with patients. 

“The 90,000 hours is a forecast of time our case managers can redirect to patient care, based on pilot data we haven’t confirmed yet,” he said. “Our goal is for every case manager to reach more patients, catch problems earlier and spend more time with the people they’re caring for.”

‘There’s just so much going on’

Wilson, Regard’s president and co-founder, shared how the technology can also help identify sudden, deadly health conditions like sepsis, an infection that can be triggered by underlying health conditions or that manifests post-surgery in some patients. 

“There’s just so much going on with the scale of data on every single patient that humans can’t keep up anymore,” he said. “There’s 50,000 data points in every patient. Imagine a 100-page book, and we have time to flip through the first three pages. When we were able to catch things earlier, we’ve seen it move (the) length of stay.”

A case study compiled by Regard amid Sentara’s pilot use of the program revealed how a doctor in their Norfolk location used the technology to solve a patient’s condition. 

Dr. Colin Findlay had suspected lack of adherence to prescribed medications was the root cause of an admitted patient’s  uncontrolled heart failure, but Regard’s comb-through of the person’s records revealed a separate condition. 

It turned out that doctors had previously paused the patient’s use of hypothyroidism medication for other reasons.  When left untreated for too long, hypothyroidism can cause rapid heart rates. Regard helped Findlay confirm that it was time for the patients’ prescription to be resumed. 

While AI use has been growing across multiple industries in recent years, it has presented a series of pros and cons. 

Theresa McDonnell, chief nurse executive at Duke University Health System, has emphasized that AI tools can reduce staff burnout, but cautioned about the technology’s potential to depersonalize care or weaken patient trust.

Likewise, a National Institutes of Health study labeled AI “the opportunity for health care revolution,” but said that it also presents ethical, regulatory and safety challenges to address when integrating it. 

Regulating the tech

In Virginia’s legislative session earlier this year, Del. Irene Shin, D-Fairfax, and Sen. Jeremy McPike, D-Prince William successfully carried legislation to prohibit downcoding insurance claims without a healthcare provider providing reasons for the decision. 

The new law is meant to ensure human decisions factored into insurance claims rather than allowing AI technology to carry out the action on its own. 

“Applications where it’s an assistive tool and not deciding is what’s appropriate,” McPike said. “We have licensed practitioners, and that’s why we hire them. That’s why we folks go to school in very sophisticated and specialized areas.”

Further action is expected in the state’s 2027 legislative session, Virginia’s Joint Committee on Technology and Science director Jodi Kuhn said. The committee outlined a series of recommendations for lawmakers to explore.

One suggestion is to establish an enforcement agency that can receive complaints from the public, investigate potential noncompliance with regulations and impose penalties. 

Other ideas include requiring disclosure of AI use in healthcare or human-involvement bills, like the ones McPike and Shin have already explored. 

Del. Rodney Willett, D-Henrico, who chairs the House of Delegates’ Health and Human Services Committee, said he will be introducing the public disclosure bill soon. 

“It’s one of those guardrail things, but it’s not prohibiting AI use,” he said.