University Health officials passed their first policy on artificial intelligence on Tuesday, providing a framework for adopting AI-powered technology across the system’s hospitals and clinics.
Artificial intelligence is already everywhere at University Health’s main hospital in the South Texas Medical Center. From assisting radiologists in detecting breast cancer and transcribing conversations between doctors and patients, to the cleaning robots that sweep the hospital’s floors, the technology is increasingly ubiquitous at not just University Health but most large hospital systems in the U.S.
“Some organizations are way ahead of others. We’re pretty far along in our AI journey,” Bill Phillips, chief operating officer at University Health, said after the Tuesday board of managers meetings. “About three years ago, we really started getting deep into it and starting to map out how we wanted to do things.”
University Health, Bexar County’s public hospital system, is undergoing a $1.7 billion expansion as it opens three new hospitals and two clinics across San Antonio. The main hospital has seen an explosion in patient volumes over the last decade.
Health care systems across the U.S. are integrating artificial intelligence in every level of their operations: direct care to patients, like assisting doctors in making diagnoses, predicting patient outcomes and personalizing treatment plans; medical devices and diagnostic technology, like CT scans at University Health that can more effectively detect strokes; assisting administrative tasks like scheduling; and a whole lot more.
Medical vendors are increasingly pushing AI-powered technology and devices, Phillips said, prompting the need for some guardrails.
“You’ve got to be able to separate what’s really applicable, what really works,” he said. “Because it’s not all equal.”
The new policy conveys the health system’s goal of using artificial intelligence to:
Reduce hospitalizations through predictive analysis and chronic disease management, improve patient monitoring to prevent complications;
Support the hospital’s workflow by decreasing administrative burdens and streamlining workflows;
Reinforce clinical judgements and decision making by doctors and other personnel.
Most importantly, though, the policy makes it clear that artificial intelligence will only be used at the hospital system to augment — not replace — clinical decisions from human health care personnel.
During the most recent Texas legislative session, lawmakers passed Senate Bill 1188, requiring health care providers to review all records created by artificial intelligence, and to disclose the use of AI to patients. The bill went into effect on Sept. 1.
In practical terms, that means that anything that comes from AI, a recommended treatment plan, diagnoses, disease indication, has to be reviewed by a human medical practitioner.
There are complex ethical and practical challenges with AI integration, including patient data privacy concerns, informed consent from patients about the use of their data in AI algorithms, and inherent biases in artificial intelligence based on the data it’s trained on, among others.
Still, health care systems largely see the technology as being transformative in bringing costs down and improving clinical outcomes.
“In this day and age, you have to really adapt to it for better outcomes,” Phillips said.