Integrating multiple transportation modes into measures of spatial food accessibility

Introduction People can access to healthy food via different modes of transportation, such as traveling by car, transit, bicycle and foot. We categorize current measures of food accessibility under an origin-destination-mode framework and find that few of them integrate multiple travel modes. As a result, these measures can bias the identification of truly low-access areas. Methods To fill this gap, we propose two new measures that integrate sub-populations of various travel modes, and estimate the overall food accessibility of a whole population. Taking Florida, USA, as a study area, we illustrate our measures with actual multiple mode commuting data from the U.S census transportation planning products (CTPP). We then compare the results to those from conventional single-modal measures. Results The proposed multiple-mode measures tend to estimate a larger population with low accessibility and fewer accessible supermarkets for a census tract, as compared to single-mode measures. The incorporation of multiple travel modes into food accessibility measures also narrows the disparities between urban and rural areas, which are indicated by conventional measures. Conclusions By considering modal-split subpopulations, our measures offer a more realistic representation of local people's travel for grocery shopping, and thus a better identification of populations with low food access. The finer modeling scale at a subpopulation level provides health and urban planners more flexibility in policy design, in that interventions can be tailored to not only a neighborhood but also a specific subpopulation within it. Such knowledge could improve the cost-effectiveness of food intervention programs.