Application of a geospatial query tool to characterize retail food environment and examine associations with dietary quality: evidence from three Chilean cities from the SALURBAL project

Background Little work has been done in Latin American contexts to characterize food retail environments and assess its influence on dietary behaviours. In particular, food retail environment data do not exist for Chilean contexts. The aim of this study is to characterize the retail food environment across neighbourhoods in three major Chilean cities and explore possible associations between the retail food environment and fruit and vegetable consumption among its residents. Methods A geospatial query tool was used to identify, georeference, and classify food establishments (FE, n = 48400) across 3 large Chilean cities (Santiago, Valparaíso, Concepción). Neighbourhood characteristics came from the 2017 Census (n = 2442 neighbourhoods) and person level data came from the 2017 National Health Survey (Encuesta Nacional de Salud, n = 1275 adults). Lower dietary quality was proxied by non-daily consumption of a fruit or vegetable (ND-FV, prevalence 17.3%). Random intercept logistic regression models adjusted for survey participant age, gender, educational attainment, neighbourhood population density and social environment. Results Median density of FE was 41.36 per km^2 and the density of FE increased with neighbourhood population density and more favourable social environment. Ready-to-consume FE and small food retail were the most prevalent types (68% and 11%, respectively); and the adjusted odds of ND-FV consumption was approximately 20% higher with increases in the z-score of those FE types. No associations were found between other types of FE and ND-FV consumption. Conclusion These findings underscore the need for targeted public policies aimed at promoting healthier food environments, reducing social inequalities in food access within Chilean cities, and improving dietary quality.