Modified Retail Food Environment Index

The mRFEI for a census tract is the percentage of food retailers in that tract that are classified as healthy — that is, healthy retailers divided by all healthy plus less-healthy retailers, times 100. A higher value means a larger share of nearby food outlets are supermarkets, large grocers, supercenters or produce stores; a lower value means the mix is dominated by convenience stores and fast food. A value of zero indicates the tract has no healthy retailers (and, if no food retailers at all, is flagged as a food desert). It is a ratio describing the retail mix, not a count of stores and not a measure of how much healthy food is actually sold.

They sit at opposite ends of a depth-versus-coverage trade-off. The mRFEI uses secondary business listings to classify outlets by type and can be computed for every census tract in the country, but it never inspects what is on the shelves. A direct-observation instrument such as NEMS sends trained raters into stores to record the actual availability, price and quality of specific healthier foods, which is far more accurate but labour-intensive and feasible only for limited samples. In practice researchers use the mRFEI for broad spatial screening and comparison and reserve store audits for validating the proxy and for detailed local studies.

Three stand out. First, it judges outlets by category, so an atypical store (a convenience store with fresh produce, or a grocery store with little) is misclassified. Second, it is only as good as the business data behind it, which contain errors and become stale. Third, census tracts may not reflect where people actually shop, especially in rural or car-dependent areas, raising the modifiable areal unit problem. Because of these, the index should be treated as a screening proxy for the retail environment, ideally ground-truthed and combined with measures of price, quality and travel before strong conclusions about food access are drawn.