CHITA 2022: Using AI to Improve Healthcare Without Introducing New Biases

Healthcare is in the midst of an AI-driven transformation, but a lot of work is needed to realize AI’s potential work in real world settings while simultaneously minimizing undesirable outcomes. From March 4-5, 2022, over 100 researchers attended the 12th annual Conference on Health IT and Analytics (CHITA) in Washington D.C., hosted by the Center for Health Information and Decision Systems (CHIDS) at the University of Maryland's Robert H. Smith School of Business. 

AI and its intersection with health disparities and bias was a central theme of the conference together with discussions on other technologies revolutionizing the practice and delivery of healthcare.

Dr. Ritu Agarwal, Distinguished University Professor, Robert H. Smith Dean’s Chair of Information Systems and Co-Director of CHIDS, commented in her opening remarks to conference participants: “...we are all bound by our collective passion for improving healthcare.  Today, the on-going global pandemic and the opportunities that digital technologies, analytics and AI offer makes the work we do more important than ever before.” 

Leading researchers and practitioners offered their insights during four keynote addresses.

“A lot of research has been conducted in AI, and a lot of papers have been published, but the translation to products and real-life change have been much slower,” said Lily Peng, MD, PhD, Director of Program Management at Google Health. “It takes more than data and models to make AI impact clinical practice,” Peng said. She then used Google’s experience in Thailand and India to illustrate the point that a “human centered approach is key to building useful products.”

Joining Peng was Dr. Griffin Weber, associate professor of medicine at Beth Israel Deaconess Medical Center and Harvard Medical School. “Data from various sources need to be linked to create a complete picture of a patient,” Weber commented in his keynote. He then explained how the federated model can enable data sharing across multiple institutions.  

Dr. Julia Adler-Milstein from the University of California at San Francisco shared her latest research on how to use the big data in electronic health records to understand physician behavior and improve clinical decision making.  

Multiple sessions at CHITA were dedicated to research on examining the possible biases that may arise from the use of AI that could further exacerbate disparities and hurt the cause of health equity. “It is critically important to ensure that AI does not introduce new biases,” commented Agarwal, the conference chair. “When we rush to apply AI, let’s remember the ‘do no harm’ principle from the Hippocratic Oath.” 

In a panel on AI and disparities, moderated by Dr. Gordon Gao, Professor and Co-Director of CHIDS, three panelists discussed important issues from practice, policy and business perspectives. “We should benchmark AI with human workers,” said Gao, “to examine whether AI has reduced systematic biases.” He was joined by Dr. Roland Thorpe of Hopkins, Dr. Jessica Galarraga from MedStar, and Mr. Cupid Chan from Pistevo Decision.

Besides AI as the central topic at CHITA 2022, telemedicine also emerged as an important issue. In a session moderated by co-chair Jeff McCullough of University of Michigan, Clarity Inc. Senior Vice President Anton Arbatov commented that “It’s about coverage and payment at the federal and state levels. We must continue to work with programs like Medicaid and Medicare to ensure [emerging] services are covered. This can be a tedious, lengthy process, because they (payers) typically require data to make decisions.” 

Telehealth policy for Medicare, for example, hadn’t changed over the past decades “largely because there was no data to prove telehealth was cost-effective,” Anton added. “We’re getting that data now. And we need to continue to push legislators and decision makers to endorse coverage as new technology rolls out. It’s an ongoing battle.”

Subsequently and relatedly, former FDA senior technology advisor Anajali Kataria told the CHITA gathering “we’re seeing convergence -- to where home becomes the epicenter of care, versus the hospital or clinic.”

“So, the doctor ‘on telemed’ will be assisting the patient and caregiver at home,” said Kataria, now CEO and co-founder of Mytonomy. “We’ll see a rise in diagnostics -- we’ll see imaging, an MRI, for example, that’s portable and arrives by Uber to your house…We’re within just years of having our own living room docking station for home health diagnostics. We’re already seeing from COVID the ability for the mass population to do lab testing – to take your own swab and do your own sample collection.”

“Healthcare is ripe for disruption,” said industry-keynote speaker Sam Hanna, pointing to accumulating data from the increasing utilization of connected health tools and remote-patient monitoring supported by wearable devices and sensors and related apps and devices – all in addition to the COVID-inspired increase in virtual doctor visits.

“The industry has had a reputation for being laggard when it comes to technology,” added Hanna, a former digital strategist for firms including PwC and Deloitte now directing American University’s master’s program in healthcare management. “But we’re now at a critical point because we have so much [actionable, consumer-generated data] toward improving the patient experience…We [now] have a way to leapfrog and do some very cool and important things that help us all.”

Thanks to the generous support of NIH’s AIM-AHEAD program, this year’s CHITA offered travel grants to 15 attendees, a majority of whom were under-represented minorities. “I see the tremendous value of AI in improving healthcare, and am eager to explore even more,” said grant recipient Dr Abiodun Otolorin, family medicine specialist and assistant professor at Howard University.

AIM-AHEAD was represented at CHITA by Dr. Legand Burge, who is leading the program’s Data Science Training Core, and Ms. Lavanya Vishwanatha, program manager.

Paper sessions and research panel discussions included: 

• AI Augmentation and Applications in Healthcare
• Big Data in Healthcare
• COVID
• Disparities, Equity and Fairness
• Electronic Health Records
• Health IT Impact
• Healthcare Biases and Disparities
• Innovations in Care Delivery
• Optimizing Resource Allocation in Healthcare
• Patient Wellness
• Physician Behavior
• Privacy
• Service Organizations in Healthcare

CHITA organizers also presented research awards:

• Best Student Authored Paper - “Convenience vs. Confidentiality: Rethinking the Role of Online Purchases at the Era of Rising Privacy Concerns” by Huiqi Zhuang, Ruli Xiao, Jian Ni and Yajing Jiang
• Best Paper Award – “Effects of Real time Prescription Benefits” by Sunita Desai, Jiejie Wang, Alan Chen, Wei-Yi Chung, Jay Stadelman, Chris Mahoney, Adam Szerencsy, Lisa Anzisi, Ateev Mehrotra and Leora Horwitz
• Young Researcher Award – Sriram Somanchi of University of Notre Dame’s Mendoza College of Business for work at the intersection of machine learning and social science 
• Health IT in Action Award (new) – “Judgmental Bot: Conversational AI in Online Mental Health Screening” by Mohammad Rahimi, Antino Kim, Sezgin Ayabakan and Alan Dennis

CHITA is produced annually in partnership with the University of Michigan School of Public Health, University of Maryland School of Public Health, and is supported in part by the Agency for Healthcare Research and Quality. CHITA 2022 also received support from a Title VI CIBE grant administered by the U.S. Department of Education.