[Household food security: comparing an alternative method to a classical one]

Establishing the performance of the US Environmental Protection Agency (EPA) household food security scale (EPSA) which is being used in Latin-America and the Caribbean, compared to a traditionally-used method (food insecurity scale) which has led to establishing food security at individual and population level. The performance of the household food security scale (EPSA) was evaluated during 2007-2008 and compared to that of the food insecurity (FI) scale based on the energy usually consumed.

Two hundred and eleven household participated in the study. The person responsible for preparing food in the home answered the EPSA questionnaire. Another household member filled in a form recording the last twenty-four hours' household consumption (R24H) (on two different occasions). The study was validated by food insecurity from R24H and supposed food security from the EPSA questionnaire.

Food insecurity by R24H was 48.8 % and 19.4 % on the EPSA. The EPSA had 16.5 % sensitivity and 77.8 % specificity. Agreement between both methods according to Cohen's Kappa was -0.06 (-0.20-0.03 CI).

Assuming equivalence of methods, the EPSA greatly underestimated household food insecurity. The EPSA results compared to those arising from the R24H were not very coherent. Some implications are discussed regarding related public policy.