Intermountain Healthcare this week announced the formation of its new Data Science and Artificial Intelligence Center of Excellence. The goal? To establish a baseline assurance for ethical standards as AI and machine learning algorithms proliferate across healthcare.

WHY IT MATTERS
The new AI and data science center convenes experts from across Intermountain and beyond, drawing on their diverse areas of expertise – data analytics, applied mathematics and statistics, computer science, behavioral sciences, econometrics, computational linguistics and clinical informatics – as well as clinicians from various specialties.

As algorithms are ever more integral to care delivery in the U.S. – Intermountain notes that there are more than 130 and counting that have been FDA-approved or cleared, designed to improve detection and treatment of breast cancer, sickle cell disease, schizophrenia and more – the aim, officials say, is to set some guidelines that can help its two-dozen hospitals and 225 clinics spot algorithmic bias, combat care disparities and help ensure "responsible and ethical AI" while also safeguarding the patient experience.

"We've developed an AI Playbook as a framework to deploy and scale human-centered AI that is transparent, equitable, ethical, and above all, ensures patient privacy," said Greg Nelson, assistant vice president for analytics services at Intermountain.

"The playbook outlines goals to establish appropriate AI governance, set validation and documentation standards, detect inherent bias, ensure data integrity and promote AI literacy among caregivers."

THE LARGER TREND
Intermountain's long history of technology innovations dates back to an early electronic health record system in the 1950s and computer-assisted clinical decision support starting in the 1970s.

More recently, AI-enabled projects at the Salt Lake City-based health system have included an e-pneumonia protocol that has saved more than 1,100 lives a year since 2015, officials note – as well as a machine learning program that can identify inpatients who are at risk of decline.

Intermountain also uses algorithms to "identify diseases and infections in their early stages, including COVID-19, and prevent hospital-readmissions," said Albert Marinez, its chief analytics officer.

Currently, any data products deployed at Intermountain follow a validation process following guidelines based on U.S. and international best practices, according to the health system, with "concepts derived from the FDA's definition of validation." They require "verification and documentation to meet predetermined specifications and quality attributes."

The aim of the new center of excellence is to broaden and fine-tune policies around AI and machine learning, to improve how they're used  to inform care delivery, officials say.

"The work of this new Center of Excellence – which brings together the right algorithms and data at the right time – enables our caregivers to care for people with the best evidence and decision-support in the moment," Dr. Mark Briesacher, chief physician executive for Intermountain, said in a statement. "People expect and deserve the best experience and care in every moment with their healthcare providers."

At the upcoming HIMSS Machine Learning & AI for Healthcare virtual event, scheduled for December 14-15, a series of sessions will explore how algorithms can be deployed more safely, effectively and equitably.

And experts from across healthcare – and outside of healthcare – will discuss some of the major policy decisions that will have to be made in the years ahead by federal agencies and private-sector companies as AI becomes ever more ubiquitous.

ON THE RECORD
"Responsible AI applications automate routine tasks and create time for providers and caregivers to listen, see and feel what patients are sharing and need. AI presents providers with data-driven insights and suggested next actions for evidence-based care plans, treatments and interventions for patients," Dr. Diego Ize-Ludlow, Intermountain's chief health information officer, said in a statement.

Twitter: @MikeMiliardHITN
Email the writer: mike.miliard@himssmedia.com
Healthcare IT News is a HIMSS publication.