Skip to main content

Analysis

Free AI Mapping Tool Helps Healthcare Systems Identify High-Risk COVID-19 Communities

By Mandy Roth  
   March 25, 2020

Jvion launches online pandemic resource that factors in social determinants of health to help pinpoint vulnerable communities and "reduce the disparity of poor outcomes."

Hospitals and healthcare systems seeking to identify communities at risk for severe COVID-19 outcomes can now locate them at the click of a mouse. The COVID Community Vulnerability Map is a free online interactive resource that pinpoints—sometimes down to the census block level—exactly where the most vulnerable potential patients are located, while simultaneously revealing social determinants of health that contribute to risk.

The tool was created by Jvion, based on the de-identified health data of 30 million patients that resides in the company's artificial intelligence core database.

"Our goal is to help communities reduce the disparity of poor outcomes," says John Showalter, MD, MSIS, chief product officer for Jvion, a metro-Atlanta company that develops AI-enabled prescriptive analytics solutions for healthcare organizations.

The map could be used, for example, to determine where to focus coronavirus intervention and outreach efforts, such as which communities might benefit from mobile testing units, Showalter says. As a user homes in on a community, the tool reveals relevant socioeconomic and environmental factors, such as lack of access to transportation or nutritious food, that put patients at greater risk. The map also denotes locations of nearby hospitals, food sources, and transportation.

Because the existing COVID-19 case data is retrospective and incomplete, Jvion examined a representative sample of two million patients from the broader database to understand the impact of parallel viruses, such as influenza. Data was further analyzed to assess cases which resulted in severe outcomes, including hospitalization, respiratory failure, organ damage, sepsis, and death. The size of the dataset minimizes the risk of bias, according to the company.

The initiative was designed to produce a tool that will help healthcare providers and organizations stem the outbreak and prioritize care for the most vulnerable patients, says Showalter.

"If you look back to [2009] and the H1N1 pandemic, there were really marked disparities in severity of illness and rates of contraction between patients who had socioeconomic disparities and those who did not," says Showalter. The Centers for Disease Control and Prevention said that the lack of understanding about the socioeconomic factors driving that risk was a barrier to appropriate planning, he explains. With COVID-19, "We are trying to use our AI approach and our understanding of socioeconomic determinants across 30 million patients to fill in those gaps for health systems and communities that are planning their response," says Showalter.

The data is being analyzed for further insights. Early findings indicate that patients who attend public sporting, concert, and entertainment events are at higher risk of exposure and have a higher chance of a severe infection, says Showalter. Public transit commutes of longer than an hour also contribute to higher risk. These findings "make total sense with the information that we're currently getting about [the need for] social distancing," he says.

While Jvion is working with existing clients to identify which specific patients are at highest risk for negative outcomes, the mapping tool is available at no cost to any person or organization who wants to use it to locate vulnerable communities.

"Helping during the pandemic aligns with our values," says Showalter. "Our mission is to reduce preventable harm. We're committed to evolve this initiative and bring forward the best insights that we can."

“Our goal is to help communities reduce the disparity of poor outcomes.”

Mandy Roth is the innovations editor at HealthLeaders.


KEY TAKEAWAYS

Hospitals and healthcare systems could use the map to determine where to focus intervention and outreach efforts.

Tool uses the de-identified health data of 30 million patients that resides in the company's artificial intelligence core database.


Get the latest on healthcare leadership in your inbox.