$22 million donation to UC San Diego Health will establish a mission control center to manage emerging AI

With artificial intelligence rapidly changing health care, UC San Diego Health is planning to treat the situation with a level of attention usually reserved for rocket launches and wildfires.

A $22 million donation from philanthropists Joan and Irwin Jacobs will help pay for a mission control center inside UCSD Health’s main La Jolla medical center to consolidate the ever-growing streams of digital information that are increasingly providing actionable information at the bedside.

Hundreds gathered in a university auditorium May 5 to listen to the latest thinking about how this technological transformation is likely to unfold, with Irwin Jacobs sitting in the front row soaking up every detail.

Jacobs, a digital communications pioneer with a doctorate in electrical engineering, said during a lunch break that it was clear during the planning stages of UC San Diego Jacobs Medical Center, the La Jolla hospital that bears his name, that the proliferation of information technology in medicine would eventually require more coordination.

“It was kind of decided, well, we’re getting all of this data but none of it’s really connected,” Jacobs said. “We need to get it into one place including not just the hospital system but also from outside, and then have a few different types of people in there who can react very quickly to what they’re seeing.”

These days, everything from bedside monitors to air-handling equipment produces endless digital information, and recent advances in artificial intelligence are showing a capacity to sift through the mountain of ones and zeros to find patterns that can spot errors and, increasingly, predict who might be about to develop a new set of symptoms.

A good example, said Dr. Christopher Longhurst, the university health system’s chief medical and digital officer, is an emergency room program that is using AI to analyze bedside and electronic health records data to predict which patients are at the greatest risk of developing sepsis, a runaway reaction to infection that can cause deadly organ failure.

“We implemented this algorithm six months ago, and in our emergency department in the last six months we’ve had the lowest observed [vs.] expected mortality and sepsis that we’ve ever seen at UC San Diego Health,” Longhurst said.

Other efforts are underway to use AI to predict which patients will develop bowel obstructions after surgery, and a remote telemonitoring program is now receiving data from the homes of more than 2,500 patients with chronic diseases.

More recently, UCSD studied systems to enable AI-enhanced recommendations for its doctors to review when responding to patients’ emails.

News

Chatbot medical responses get good reviews in UCSD study

ChatGPT often answered routine questions more completely and with more empathy than busy human doctors, according to evaluators.

And this is just the beginning. Every new application, Longhurst said, will generate its own set of notices. Asking bedside workers to parse this flow is impossible, meaning that a separate team of professionals will be necessary to decide what needs to be passed along to caregivers and what can wait.

Though the need for such an approach is already arriving, Longhurst said it is expected to take several years for the new command center, which also will have two components outside the hospital, to get up and running.

For now, AI’s potential to make routine tasks more manageable for medical professionals seems to be creating the most excitement.

Panelists who spoke during the symposium were asked what excites them the most in the near term about the coming AI health care revolution. Most said they were optimistic about the ability of algorithms to help free up medical professionals’ time by assisting with routine tasks, such as responding to patients’ emails for medical testing and other routine communication that piles up during the workday and impinges on personal lives. Having help grinding through the grist of modern health care should, in theory, free up time for meaningful conversations with patients.

Holly Smith, UCSD’s population health clinical nurse educator, said she hopes that’s how it goes. Technology, she said, will clearly be necessary to meet the ever-increasing health care demands of an aging population.

But gains in efficiency must be balanced with the human touch, she said.

“We don’t want to use that data to make things so efficient that we take that human piece out,” Smith said. “A careful, managed approach is important, and so is making sure to get the perspective of the people that are in the field doing that frontline work with the patients.” ◆