Poster #68 - Integrating Environmental Barriers into Food Desert Maps: Robustness and Real-world Relevance

• In Event: Saturday Poster Luncheon

Abstract

Current food desert definitions mainly use distance to supermarkets as a proxy for food access. In urban environments, residents rely heavily on active transportation (e.g., walking). Compared to healthy young adults, older adults are more susceptible to experiencing environmental factors as barriers to active transportation. An older adult, for example, is more susceptible to heat stroke while walking an apparently short distance to get the groceries they need, if the journey is along a road with no tree shade on a hot summer day. Current methods likely overestimate food access under these conditions.

Previously, we implemented a new approach to defining food deserts that accounts for environmental barriers. We focused on the elevated risk of heat stroke among older adults in summer weather, incorporating the impact of temperature, humidity, and heat index environmental variables to calculate a metric we call “Real Feel” distance. Calculating this metric for a single hot summer day in Chicago, we found that the number of older adults that live in census tracts where the average distance to a grocery store is un-walkable may go from 150K (44%, out of 340K total) under ideal environmental conditions to 250K (an additional 30%) on a hot summer day. This increase is substantial, and points to a large overestimation of food access in this example implementation

Here, we test the robustness and real-world relevance of this result.  

To test robustness of our result we 1) extend calculation of the Real Feel metric and its impact from a single day to across a full calendar year to check the annual trend for consistency and expected pattern, and 2) test sensitivity of the food access overestimation result to parameterization and formulation of the metric calculation. We note that most people expect to shop for food at least once a week. To test real-world relevance, we test for cases where Real Feel access is low for full week stretches, interfering with this expectation.

Annual Real Feel metric trends, as expected, show impact on access isolated to summer months, rising in June to a peak in July/August, and falling again in September. Consistency is high, with tight confidence intervals.

Metric calculation includes terms for each environmental variable and age. Sensitivity tests included 1) dropping potentially redundant temp and humidity terms, 2) additionally, or only, dropping potentially conflating age term, 3) widely varying temp, humidity, and heat index term parameters. Across variations, minimum access overestimation was 5%; Mean overestimation was 15% +/-6%.

Finally, one third of Chicago census tracts experienced at least one full week stretch of low Real Feel access during the year.

Given these preliminary results, the potential is high for extending this new approach to defining food deserts. Here, we assessed food access for older adults in Chicago. We can generalize this approach to assess access to other resources, for other populations, in other cities. Interactive tools incorporating this approach — also with increased geographic and temporal resolution — may prove useful for policymakers, and for residents making daily destination and route decisions.

Author

• Andrea Tentner, NORC at the University of Chicago
  Presenting Author