The machines arrive in rural Washington: How AI is helping improve health equity
What most people know about Washington comes from its portrayal in TV and movies – rainy skies, towering trees, stunning mountains ranges, coffee drinkers in flannel shirts. But what many don’t realize is that much of the state is rural, including the Okanogan County region served by Three Rivers Hospital in Brewster. There, a large portion of the local economy is driven by agriculture, and much of the labor is provided by Hispanic farmworkers.
The local Hispanic community faces its own challenges when it comes to accessing health care. In addition to facing more common barriers to care like limited health services and transportation access, rural Hispanic farmworkers also grapple with language barriers, lack of health insurance, and experience income-related challenges tied to minimal labor protections.
To better understand this community’s health care needs, Three Rivers Hospital is leveraging Washington’s renowned skill in technology to deploy artificial intelligence and machine learning in service of identifying and eliminating health disparities.
“I am fascinated by the amount of sociological factors underlying health outcomes and their disparities. For example, many older patients in rural areas do not have SMS-enabled cell phones and rarely interact with online patient portals,” said Washington State University Sociologist and Assistant Professor Anna Zamora-Kapoor, a fellow in the Artificial Intelligence/Machine Learning Consortium to Advance Health Equity and Researcher Diversity (AIM-AHEAD) leadership program. The program, launched in 2021, is a joint effort between Three Rivers and the National Institutes of Health.
One focus is on improving cancer survival outcomes among Hispanic patients, especially that of the lungs, which is the leading cause of death in this group. AI-generated text images and text-based intervention are used to help identify, schedule and follow up with patients who are eligible for a low-dose computed tomography scan, which screens for lung cancer.
Another focus is on increasing participation and representation of researchers and communities currently underrepresented in the development of AI/machine learning models and to enhance these models’ capabilities to address health care disparities.
Hispanic people are one of the groups underrepresented in AI/ML models. This underrepresentation of certain groups can have major implications for health equity. As model data sets become more diverse through initiatives such as this, the likelihood of harmful biases being created and perpetuated decreases.