Cost-Effective Obesity Surveillance with DMV Data

November 11, 2015

When Mayor Francis Slay included obesity reduction (five percent by 2018) as a key goal in his Sustainability Action Agenda for St. Louis, the city’s Department of Health (DOH) enlisted Ben Cooper, manager of the Institute for Public Health at Washington University’s Public Health Data and Training Center for help collecting baseline data.

Often this type of information comes from the CDC’s Behavioral Risk Factor Surveillance System (BRFSS), but the most recent dataset for St. Louis was three years old and did not allow for geographical analysis more specific than county level. To best identify target populations for obesity intervention initiatives, ward- or neighborhood-level data is essential. Cooper learned of a project led by Daniel S. Morris in Oregon that performed obesity tracking through driver’s license data, and decided to attempt a similar effort in St. Louis. He connected with Carl Filler and Matt Steiner in the DOH, and the three worked together on this project.

Filler and Steiner requested records from the DMV, and they received information on 391,000 individuals in the metropolitan St. Louis area, including gender, birthdate, street address, height, and weight. When data was cleaned and limited to people residing within the City of St. Louis, the team had almost 172,000 individual records which could be plotted on a map. Then analysis was conducted to determine the percentage of overweight/obese people by zip code, census tract, and neighborhood. Steiner created the maps below, and the data and its analysis provided key information for the 2015 City of St. Louis Obesity Report.

The results of the study showed that roughly 60.9 percent of City residents included were overweight or obese and fewer than two in five (38.3 percent) were normal weight. Young people overall had lower BMIs, with only about 33 percent of residents under 20 years old being overweight or obese compared to 72 percent among 50-79 year olds. Geographical analysis found the highest obesity rates present primarily among neighborhoods in northern St. Louis, with most areas reporting over 70 percent of their populations to be overweight or obese. (North St. Louis is also impacted by higher unemployment, poverty, and crime rates than the rest of the city). This indicates that obesity intervention programs targeted to these neighborhoods could potentially have the most significant impact on overall public health in the city.

Although not collected for research purposes, Cooper believes DMV data can be a useful and inexpensive tool for public health surveillance, especially related to obesity. He hopes future research will integrate geospatial imaging with BMI data to examine the relationship between obesity rates and access to mass transit, parks, grocery stores, and other social and environmental resources in St. Louis, as well as health outcomes.