How AI Is Helping More Patients Heal at Home

by Chris Nichols
| Nov 16, 2026

The Big Picture

Hospitals across the country face capacity challenges, even as some patients could safely receive hospital-level care at home. University of Utah Health’s Heal at Home program is helping bridge that gap—and an AI-supported tool is making it easier for clinicians to identify patients who may be ready to recover at home.

A partnership between U of U Health and Community Nursing Services, Heal at Home connects qualifying patients with hospital-level services at home while keeping them connected to their care team. Since launching with two specialties in 2021, the program has grown to include 22 specialties and has served more than 1,500 patients.

What’s Happening

Finding the right patients at the right time has been a challenge. For hospital medicine patients, clinicians previously reviewed a daily inpatient census that could include 115 to 160 patients, looking for people who might qualify for home-based care. The process was time-consuming and increasingly difficult to scale.

To help, clinicians, developers, and AI specialists worked together to build a tool that evaluates patient information, including vital signs, laboratory values, clinical stability indicators, and exclusion criteria. It surfaces patients who may be appropriate for Heal at Home, giving the team a smaller group to review and helping them start conversations with care teams earlier.

The tool is designed to account for clinical differences across conditions: what stability looks like for a patient with heart failure may differ from what it looks like for someone recovering from acute kidney injury. Patient data remains within U of U Health’s secure infrastructure.

Why It Matters

Timing is critical when a patient is approaching discharge. Identifying potential candidates sooner gives clinicians more opportunity to discuss whether home-based care is appropriate before that window closes. It also reduces the time spent manually reviewing charts, freeing clinicians to focus on patient care.

Heal at Home has reported fewer emergency department visits and readmissions compared with traditional home health discharges, along with patient satisfaction rates near 99%. These are program outcomes, not measures of the AI tool’s individual impact.

The Bottom Line

The AI tool supports clinicians; it does not make the final decision about a patient’s care. By bringing relevant information forward, it helps the care team have a human-to-human conversation about whether Heal at Home is right for each patient at that moment—and could help the program grow without a corresponding increase in administrative burden.