Sponsored

AI and triage redesign: Inside Inspira Health’s emergency department transformation

Advertisement

Emergency departments nationwide face mounting pressure, from rising patient acuity to staffing shortages and operational inefficiencies.

In a recent discussion hosted by Becker’s Healthcare and Mednition, leaders from Inspira Health, a four-hospital system in southern New Jersey, shared how they are tackling these challenges by integrating AI into ED triage and redesigning key clinical roles.

Here are four key takeaways from the session:

1. High under-triage rates sparked action

Inspira leaders proactively identified a significant opportunity to improve emergency triage accuracy, a widespread challenge that emergency departments nationwide are grappling with. Pre-implementation data revealed a 51.1% under-triage rate for high-risk patient presentations and a 23.2% under-triage rate across all emergency department visits.

“This data is one of the primary reasons we decided to explore technology such as KATE AI to support our nurses,” said Kate Mangarelli, assistant vice president of emergency and nursing operational services for Inspira Health.

By embedding the AI tool into the existing EHR-based triage workflow, Inspira avoided process disruption. The system provides real-time notifications to the nurses when documentation and ESI (Emergency Severity Index) scores don’t align with patient risk.

2. AI improves accuracy and documentation

After rolling out Mednition’s KATE AI in mid-2024, Inspira saw a sharp improvement in predictive accuracy and nurse documentation. The AI-powered triage predictions now appear on more than 90% of ED records, up from 80% before implementation.

More importantly, under-triage rates for high-risk patients fell from 51.1% to 39.9%.

“We have an algorithm running concurrently making early identification easier, and supporting early intervention and bundle compliance,” said Deena Brecher, chief nursing officer at Mednition. “It’s essential to remember that this is a nurse-first approach, meaning it’s meant to support or enhance and complement the nurse, not replace them.”

3. Developing custom nurse roles

Alongside AI deployment, Inspira introduced two new roles in the ED: the flow nurse and the resuscitation nurse. The flow nurse oversees patient movement and capacity across the department, while the resuscitation nurse focuses exclusively on high-acuity care and rapid stabilization.

“We recognize the need to develop solution-based workflows which enhance the timeliness of care, the development and implementation of the flow and recess nurse roles supported our goal of getting the patient to the right place at the right time,” said Sherry Remy, assistant vice president of Emergency & Nursing Operational Services at Inspira Health.

The results include measurable improvements: door-to-EKG times for chest pain patients improved by 10 percentage points overall, and door-to-needle times for stroke patients improved by six minutes.

4. Change management matters with AI

Initial skepticism among nurses gave way to widespread support as they saw the value of real-time AI notifications in their triage assessments. The team credits this to a culture of transparency and the leadership’s decision to position KATE as a supportive tool rather than a punitive measure.

“Nurses are pretty data-driven, so I think it’s helpful when you share the data with the team members because they like to know how they’re doing and we’ve definitely seen more adoption,” said Brooke Zacheis, vice president of clinical services at Inspira Health. “It was challenging, but now they’ve embraced it, they love it and look at it as a resource.”

Inspira also uses nurse-specific performance dashboards to inform targeted education and build triage confidence, particularly for newer clinicians.

With KATE AI now embedded in all ED locations, Inspira is exploring further applications including predictive alerts for decompensation and clinical data integration in areas such as maternal health and heart failure. They’re also working closely with Mednition to expand AI use beyond initial triage to support real-time reassessment throughout the ED stay.

“We no longer have the luxury of having seasoned nurses in triage every day, ” Ms. Zacheis said. “Having a tool like this gives our leaders and our educators insight into taking a more proactive approach and using data to recognize who may need additional triage support with the fundamentals. This data not only drives future training strategies, but it emphasizes the importance of continued learning and engagement with our nurses.”

Advertisement

Next Up in Artificial Intelligence

Advertisement