Can Computer Vision Models Help Us Picture Better Health Outcomes?

Guest post by Andrew Wiley, Video Producer, NLM Office of Communications and Public Liaison (OCPL).

NLM recently featured a video that highlights an exciting project funded through the NLM Extramural Programs Division.

A team led by Quynh Nguyen, PhD, MSPH, Assistant Professor of Epidemiology and Biostatistics at the University of Maryland School of Public Health, used a massive repository of millions of images of all road intersections and street segments in the United States to characterize neighborhood-built environments. Neighborhood attributes have been shown to influence health; however, advances in neighborhood research have been constrained by the lack of neighborhood data for many geographical areas, and few neighborhood studies examine features unique to rural locations.

Dr. Nguyen and her team used regression models to estimate associations between features of built environments and county health outcomes, controlling for risk factors like county-level demographics, economics, and population density. Use of health information allowed researchers to determine relationships between the characteristics of a neighborhood and individual-level health outcomes. Findings from this research indicate that built environment features such as a lack of crosswalks in neighborhoods, housing density, and single-lane roads may be associated with a variety of diseases (e.g., diabetes, cardiovascular disease, etc.) as well as adverse health outcomes including premature mortality.

Watch the video about this research and find the transcript here.

Get More Video Content and Learn More About NLM

Visit Welcome to NLM to take a virtual tour and to learn more about NLM’s programs, tools, and resources. Be sure to subscribe to the NLM YouTube channel for access to more great video content.