NIH Grant Aims to Enhance Artificial Intelligence Use in Healthcare

After receiving $5.9 million in funding from the National Institutes of Health (NIH), two researchers from the University of Virginia (UVA) aim to determine how artificial intelligence (AI) can support care for diverse populations.

Although AI use in healthcare is growing in terms of technology innovation and application, its development is often based on non-diverse populations, limiting its widespread use as it could harm non-White patients.

The NIH aims to learn about the patterns of living systems and use this information to enhance life and health through various efforts. The press release also indicates that the organization aims to eliminate healthcare disparities through its “Bridge to AI” programs, which involve a group of scholars seeking to improve healthcare AI development and use.

Following a grant from the NIH, primary co-investigators Ishan Williams, PhD, an associate professor in the UVA School of Nursing, and Randall Moorman, MD, a UVA Health cardiologist, will lead an effort to create and test AI to enhance healthcare for more diverse populations. Moorman is also the Bicentennial Professor of Advanced Medical Analytics in the school of medicine and has extensive experience with AI.

According to the press release, COVID-19 had a larger, more negative impact on communities of color compared to those who are White. A significant focus of this project will be to consider race, ethnicity, socioeconomic status, and geography data when developing and testing AI solutions.

“If COVID has taught us anything, it’s that we must take a nuanced, inclusive, equitable approach to everything we do in health care, AI included,” said Williams in a press release.

Moorman and Williams work with investigators from Massachusetts General Hospital, the University of Florida, and other colleagues across the country.

This research effort began in late 2022 and is expected to be completed by 2027.

Research surrounding the inequities that derive from AI is a growing field of study.

In April 2022, researchers from PLOS digital health found that health biases and data gaps in AI solutions can often lead to care disparities.

They arrived at this conclusion following a review of studies from the US and China, which included over 7,000 articles. Researchers noted that these studies had an overrepresentation of the radiology specialty, a high variation in author demographics, and a lack of consistency in article origins. Each of these indicates a potential to affect data and create biases.

As a result of the disparities that may derive from AI use, there have been many efforts to eliminate bias.

For example, in March 2022, Dascena developed a non-biased machine-learning algorithm to enhance treatment for acute coronary syndrome. This effort occurred with the support of a grant from the National Institute on Minority Health and Health Disparities.

The machine-learning algorithm will review EHR data to determine the likely causes of bias while maintaining the aspects of the data that lead to accurate measurements. Researchers also intend for it to be trained on preprocessed data to assess its performance across subgroups.