To determine whether neighbourhood supermarket and convenience store availability and broader built environment context are associated with food purchasing behaviour in a national population.

We used observational data to perform a cross-sectional study of food purchases for US households in 2010. We used three-level mixed-effect regression models to determine whether the associations between the number of neighbourhood supermarkets and convenience stores and the self-reported annual household expenditures for fruits and vegetables were affected by regional destination accessibility, neighbourhood destination diversity, availability of neighbourhood destinations and neighbourhood street connectivity.

Metropolitan statistical areas (n 378) in the USA.

Households (n 22 448).

When we controlled for broader built environment context, there was no significant association between availability of neighbourhood supermarkets and expenditures on fruits and vegetables; instead, we observed an inverse association between the number of convenience stores and expenditures for fruits (P = 0·001). The broader built environment context was associated with food purchase, although the magnitude was small: (i) higher regional destination accessibility was associated with higher expenditures for fruits (P < 0·001); (ii) higher neighbourhood destination diversity was associated with lower expenditures for vegetables (P = 0·002); and (iii) higher neighbourhood street connectivity was associated with higher expenditures for fruits (P < 0·001).

The broader built environment factors contributed to understanding how people use neighbourhood food stores. However, there was only a small relationship between the broader environment context and fruit and vegetable expenditures. Policy interventions that focus exclusively on increasing the availability of neighbourhood supermarkets likely will not promote fruit and vegetable consumption.