Clinical AI gains ground in a resource-constrained hospital

Operational pressures drive AI investment decisions

Like many regional hospitals, San Juan Regional's challenge was not a lack of clinical data but the difficulty of using it efficiently. Hallak said clinicians often navigate multiple screens and external resources to find necessary information, which can delay care decisions and affect patient flow.

Hallak described the longstanding challenge facing IT leaders: "How can we present the right data at the right moment to the right level of care?" he said.

That challenge ultimately led the hospital to adopt an AI-based clinical workflow platform, starting in January 2026.

Hallak emphasized the hospital's investment was not driven by excitement around AI itself.

"It is not a hype," he said. "It has solved a problem that we've been trying to solve for many years as clinicians."

San Juan Regional is not alone in exploring this approach. In a press release, Wellsheet reported that its platform is now deployed in more than 100 hospitals nationwide, reflecting growing interest in workflow-integrated clinical AI tools.

Fitting in with existing workflows

Hallak said his team evaluated potential tools based largely on how well they would support clinicians within their existing workflows.

According to Hallak, the AI software's integration with UpToDate -- a widely used, evidence-based clinical decision support resource relied on by clinicians for current research and treatment guidance -- became a major deciding factor during the selection process.

Traditionally, providers consult such references separately from the EHR, manually applying patient data to clinical guidance. Wellsheet streamlines that workflow by connecting UpToDate with the EHR.

By presenting guidance in the context of an individual patient, clinicians can review relevant evidence without leaving the patient's chart.

"Physicians don't have to get outside of the EHR and read an article that is not in the context of the patient," Hallak said. "It's based on that patient and their information."

The platform summarizes chart information and links clinical pathways, decision-support calculators and other reference material using patient-specific information from the medical record. Rather than directing treatment, the system organizes information, Hallak emphasized.

It's an approach that builds on traditional clinical decision support models, in which evidence informs care while responsibility remains with the clinician.

"I'm giving you the right information to make the right decisions," Hallak said.

Patient outcomes, clinician satisfaction will indicate success

San Juan Regional plans to evaluate the investment using operational and clinical metrics, including discharge efficiency and length of stay. Hallak said they also hope to see earlier identification of patient deterioration.

"I'm looking for better clinical and operational efficiency -- being able to discharge the patient at the right time safely," Hallak said. "Hopefully, in six months, down the road, we will report great outcomes."

He said the hospital is also paying attention to how the technology affects day-to-day clinician experience. By reducing the need to search through multiple screens for information, leaders hope the platform will make workflows feel more manageable.

"It's about physician and care team satisfaction -- not struggling to search for data, but bringing back some joy to practicing medicine and using technology," Hallak said.

San Juan Regional's experience reflects a broader shift among community hospitals seeking operational efficiency amid mounting pressures. Ultimately, health systems with limited resources will have to judge clinical AI by its ability to improve efficiency and patient care alongside the cost of the technology.

Elizabeth Stricker, BSN, RN, comes from a nursing and healthcare leadership background, and covers health technology and leadership trends for B2B audiences.