In the USA, community-based food pantries provide free groceries to people struggling with food insecurity. Many pantries obtain food from regional food banks using an online shopping platform. A food bank introduced a visible nutrition rank (i.e. green, yellow or red) onto its platform. The hypothesis was that pantry orders would increase for the healthiest options (green) and decrease for the least healthy options (red).

Interrupted time series (ITS) analysis of a natural experiment. Monthly data included nutrition ranks of available inventory and itemised records of all products ordered during the 15-month baseline period and 14-month intervention.

A New England food bank.

The twenty-five largest food pantries in the network based on pounds of food ordered.

Descriptive analyses of 63 922 pantry ordering records before and after the visible ranks identified an increase in the proportion of green items ordered (39·3–45·4 %) and a decrease in the proportion of red items ordered (10·5–5·1 %). ITS analyses controlling for monthly changes in inventory available and pantry variables indicated that average monthly orders of green items increased by 1286 pounds (P < 0·001) and red orders decreased by 631 pounds (P = 0·045). Among the largest changes were increases in orders of fresh produce, brown rice, low-fat dairy and low-fat meats and decreases in orders of sugary juice drinks, canned fruit with added sugar, higher fat dairy and higher fat meats.

This promising practice can support system-wide efforts to promote healthier foods within the food banking network.