Back
Science

Researchers Develop Machine Learning Index to Map U.S. State Flu Vulnerability

View source

New Flu Vulnerability Index Highlights U.S. Regional Disparities

Researchers at Washington University in St. Louis have developed a new state-level vulnerability index for assessing the risk of influenza-like illness. The findings, published in PLOS Computational Biology, highlight significant regional disparities in infection risk across U.S. states.

Advanced Methodology Uncovers Complex Risk Factors

Unlike previous models that primarily relied on health data, this new approach uses machine learning algorithms. It integrates 39 socioeconomic and health indicators from census data, including factors such as migration patterns, insurance coverage, and the proportions of female and elderly populations. This methodology aims to determine non-linear relationships between these factors and vulnerability to flu.

Identifying Vulnerability Hotspots for Targeted Strategies

Rajan Chakrabarty, a professor at Washington University’s McKelvey School of Engineering, emphasized the index's utility:

"The index can identify vulnerability 'hotspots' in the U.S., enabling policymakers to develop more targeted strategies for flu prevention."

Shrabani Sailaja Tripathy, the lead author and postdoctoral associate in Chakrabarty's lab, further elaborated on the nature of vulnerability:

"Vulnerability is a result of multiple factors such as urbanization, demographics, healthcare access, and economic disparities."

This comprehensive approach aims to provide a clearer picture for targeted intervention.

Unique State-Level Risk Profiles Emerge

The research indicates that each state possesses unique risk factor profiles.

For instance, the District of Columbia exhibits high risk due to population density, mobility, a significant uninsured foreign-born population, and longer commute times.

States like New Mexico and Arizona, with large rural regions, show heightened flu vulnerability linked to higher numbers of aging, female, and Hispanic people, who are at increased risk for flu complications.

Michigan is identified as another high-risk state, experiencing dual challenges from high transmission risks in urban areas and economic hardships in rural regions.

A Tool for Broader Epidemic Preparedness

The researchers believe this new tool can be applied to understand vulnerability to any infectious disease, thereby strengthening epidemic preparedness and response.