AI helps save lives at UC San Diego Health

UC San Diego Health is on the forefront of integrating artificial intelligence into clinical and operational processes, and it's beginning to save lives.

The health system uses an AI algorithm called COMPOSER to predict sepsis before clinical manifestations, according to a study published in January. Patients who check into the emergency department are continuously monitored for more than 150 variables that could be linked to sepsis, including lab results, vital signs, medications and more. When patients are at risk, the AI algorithm notifies nursing staff through the EHR.

"We found early detection of sepsis resulting in 20% drop of mortality in our emergency department patients with sepsis, which is really, really huge and translates to 50 lives saved on an annual basis here at UC San Diego," said Chris Longhurst, MD, chief medical officer and chief digital officer at UC San Diego Health, said on an episode of the "Becker's Healthcare Podcast." "Imagine if that translated across the 300,000 lives lost every year across America."

UC San Diego has also leveraged large language models for AI generated draft messages to respond to patients on the portal, which decreased cognitive burden on clinicians and lowered physician burnout.

The health system is seeing results with AI applications on operational functions, including making the revenue cycle more efficient. Now when UC San Diego acquires new hospitals, they're able to expand services without having to add as many people as before.

"That efficiency also translates to better employee satisfaction," said Dr. Longhurst. "We know that some of the rote tasks that can be done by AI, when we automate that, makes our employees happier and leads to increased retention and overall a better healthcare experience."

As AI becomes more integrated into healthcare, Dr. Longhurst recommends leaders have at least a serviceable understanding of how it works and possibilities for the future.

"As AI begins to really creep into every corner of healthcare delivery, having a superficial understanding of how these tools work is actually important for some of our executive leaders," said Dr. Longhurst. "It's fantastic to move forward to identify where the opportunities are to leverage these tools to drive the clinical and business outcomes that matter to our patients and to the organization."

UC San Diego Health has a large Medicare Shared Savings Program accountable care organization and they've built a remote patient monitoring program using devices at home to transmit blood pressure and glucose readings for diabetic patients to their care teams. There are around 3,000 active patients in the program, which has helped better monitor the diabetic population.

"We actually have data showing a reduction in blood pressure in our hypertensive patients, a reduction in hemoglobin A1C of our patients with diabetes," said Dr. Longhurst. "That kind of care allows us to not only drive better outcomes and reduce hospital admissions, but also saves costs to the system overall and keeps people healthy and out of the hospital."