Orlando, Florida – Artificial intelligence emerged as a key theme at the HIMSS Global Health Conference & Exhibition, where leaders discussed tools to detect disease earlier and improve outcomes.
Just as many in the healthcare industry are excited about the potential of AI, healthcare leaders are also concerned that advances in AI may not be available to everyone. Many have warned that if AI is not developed with health equity as a guide, it could exacerbate disparities among disadvantaged communities.
Jennifer Stoll, OCHIN's chief external affairs officer, says she is very concerned about the potential for AI to lead to increased inequalities in health care. He OCHIN, a nonprofit organization, works with federally qualified health centers and local hospitals to improve technology capabilities.
“This is going to be the biggest challenge with AI,” Stoll told Chief Healthcare Executives® in an interview at the HIMSS conference. “If not done carefully, you end up with a unique set of haves and have-nots.”
OCHIN has partnered with Microsoft and more than a dozen major hospital systems to form the Trustworthy & Responsible AI Network, also known as TRAIN. OCHIN wants to ultimately make these AI tools accessible to providers with fewer resources in underserved communities.
“AI has the potential to drive tremendous efficiencies, improve outcomes, and be a great source of knowledge for those serving rural and underserved communities. Yes,” Stoll said. “But if not carefully and thoughtfully deployed, it can also become a great destroyer of the world.”
(In this video from the HIMSS conference, healthcare leaders talk about AI and health equity. Story continues below.)
AI can help “if done right”
Hackensack Meridian Health CEO Robert Garrett, in his keynote address at last week's HIMSS conference, talked about the potential of AI to ultimately improve the health of billions of people around the world. talked about. Garrett said that in health care, AI has the potential to improve not only outcomes but also access and equity, but he also included an important condition:
“There is no question in my mind that if done correctly, AI can help advance health equity,” Garrett said.
“I think about health equity in terms of identifying the social determinants of health: who is at risk for one or more of the social determinants of health? And we're thinking about how we can actually address those at-risk populations by connecting them to good care and medical care.''It's a great resource,'' he added. “AI can really fill that gap and hopefully close some of the disparities that exist in medical outcomes today.”
Anna Schoenbaum, vice president of applications, predictive health, and digital health at Penn Medicine, emphasized the importance of considering health equity in the development of AI tools.
“It is our responsibility to ensure that AI includes health equity,” she said in an interview at the HIMSS conference. “I think this is a huge AI tool. We need to make sure it's comprehensive.”
At the University of Pennsylvania, Schoenbaum notes that while researchers are considering the use of AI in predictive models, they are also carefully examining the data.
Health leaders say AI tools rely on data accuracy, so if they use inaccurate data or data that reflects bias against racial groups, inequalities will continue or even widen. He emphasizes that it is possible. The World Health Organization calls for caution in the use of AI tools in medical decision-making.
Researchers found that chatbots provided answers that reflected racial bias. digital medicine Last October. The Coalition to End Racial Discrimination in Clinical Algorithms found that testing bias in areas such as kidney function delays black patients from receiving better treatment. The coalition's efforts are creating changes in hospitals and health care systems.
“A mirror that reflects humanity”
Heather Lane, senior architect at Athena Health, said in an interview at the HIMSS conference that the healthcare industry needs to take “intentional action” to ensure advances in AI do not worsen health equity. Ta.
“There are certainly examples of how AI can be used to correct human biases, if done carefully, which is great,” Lane said. “But if you do it naively, you just end up absorbing all of our biases. I often say that today's AI is in many ways a mirror. It's a mirror of humanity. . And it reflects our good selves, but it also reflects our worst selves.”
Lane said health systems cannot innocently build AI solutions that deepen inequalities. She said the easy and wrong approach is to simply use existing data that may contain inequities in the system.
“If we don't think about it, and if we don't approach it carefully with measurement, safety, and equity in mind, we're going to create a system that only increases inequality,” Lane said.
“But in the right way, with the right metrics, and with careful and thoughtful optimization, you can actually reduce those inequalities, creating a system that is fairer than your initial input data.” “You can,” she added. . “And I think that’s the goal we should all strive for.”
Brendan Watkins, chief analyst at Stanford Medicine Children's Health, said there is a risk of bias being embedded in algorithms. He said he is encouraged by the many conversations about AI and health equity.
“The people I talk to at least have this in mind,” Watkins said. “So I think that’s a really good thing.”
Schoenbaum said AI can leverage data on social determinants of health collected by health systems and share that information among other health systems to provide better care and improve patient outcomes. I pointed out that it can be done.
“We think this allows us to address and advance health equity in our health solutions,” Schoenbaum said.
TRAIN's hope
With the introduction of TRAIN, Stoll said OCHIN will be able to ingest more data from disadvantaged populations into its AI, including data on Black and Hispanic patients and chronically homeless patients.
Stoll is excited about the potential for organizations with limited resources to reliably benefit from AI. She points out that smaller hospitals don't have the talent, resources, or governance to develop their own AI tools, but they need shared experience.
While some smaller hospitals and healthcare providers may be a long way from using AI tools to support clinical decision-making, they could find welcome support from tools that support business operations. Stoll said there is a possibility that the Many resource-strapped healthcare providers are turning to AI to ease the burden on clinicians and staff with tools to easily submit claims and summarize patient conversations. You will receive maximum benefit.
“We believe our community is well-positioned to take full advantage of operational efficiencies, simplified management, and reduced burden on providers,” Stoll said.
“Where OCHIN will slow down a little bit is in the clinical area, and we'll look at it very slowly from the research side,” Stoll said. “But in terms of operational efficiency, we're moving forward very quickly in this area. But it can create haves and have-nots, and that's especially true based on where we are. It only accelerates health inequities.”
While OCHIN aims to be a voice for health equity in the TRAIN initiative, Stoll also hopes to involve other organizations serving underprivileged communities.
“I think OCHIN really wants to make sure that other voices are included to support this equity conversation,” Stoll said. “I mean, we support so many people, and we're growing every day. But we need more communities to get involved.”
The federal government can also play an important role in providing funding to help smaller hospitals and providers develop AI capabilities. Stoll noted that small health care providers have received little support in transitioning to electronic health records.
“We must continue to invest in health information technology and the infrastructure needed in rural and underserved communities,” Stoll said. “That's going to be the biggest problem that leaves people behind. If we don't help everyone reach a level playing field, we're going to see more inequality.”
