Relationship between community characteristics and impact of calorie labeling on fast-food purchases.

Abstract

The objective of this study was to evaluate potential sources of heterogeneity in the effect of calorie labeling on fast-food purchases among restaurants located in areas with different neighborhood characteristics. In a quasi-experimental design, using transaction data from 2329 Taco Bell restaurants across the United States between 2008 and 2014, we estimated the relationships of census tract-level income, racial and ethnic composition, and urbanicity with the impacts of calorie labeling on calories purchased per transaction. Calorie labeling led to small, absolute reductions in calories purchased across all population subgroups, ranging between -9.3 calories (95% CI: -18.7 to 0.0) and -37.6 calories (95% CI: -41.6 to -33.7) 2 years after labeling implementation. We observed the largest difference in the effect of calorie labeling between restaurants located in rural compared with those located in high-density urban census tracts 2 years after implementation, with the effect of calorie labeling being three times larger in urban areas. Fast-food calorie labeling led to small reductions in calories purchased across all population subgroups except for rural census tracts, with some subgroups experiencing a greater benefit.