The Food Environment Toolbox: Developing and Piloting a Suite of Tools to Measure Food Environments in Low- and...

Measuring food environments in low- and middle-income countries (LMICs) can inform and support the design of interventions aimed at improving diets and reducing malnutrition in all its forms. Most food environment measurement tools have been developed for high-income countries, however, and do not sufficiently capture the diverse and dynamic food environments in LMICs.

The objective of this study was to develop, modify, implement, and refine a suite of tools to measure the different dimensions of diverse food environments in LMICs.

Through an iterative process, we: 1) identified the food environment dimensions to be measured; 2) identified existing tools using literature searches; 3) modified and developed tools based on feedback from team members of study settings, an expert advisory board, and workshops with key food environment experts in India and Cambodia; 4) implemented the tools by piloting the suite in rural, peri-urban, and urban settings in India and Cambodia; and 5) finalized tools based on feedback from experts and our tool piloting implementation and analysis experience.

Overall, we included 7 tools in the finalized Toolbox (Participatory Mapping, Seasonal Food Availability Calendar, Food Environment Perceptions Survey, Community Food Environment Mapping, Market Mapping, In-depth Vendor Assessment, and a Cost of a Healthy Diet data collection protocol), all of which were rated positively by workshop participants. On the basis of piloting experiences, the tools were relatively easy to implement in the field. Apart from the Seasonal Food Availability Calendar being better suited to rural or peri-urban settings and the In-depth Vendor Assessment being less suitable for large formal supermarkets, we found that the tools were feasible and useful across pilot settings in India and Cambodia.

The suite of tools included in the Food Environment Toolbox can be used to measure diverse food environments in LMICs, with minimal anticipated adaptations across contexts.