To analyse usual intakes of energy, macronutrients and micronutrients, and their percentage of inadequacy, in a Brazilian population at severe food insecurity (SFI) risk, determined from a predictive model using two national databases.

Cross-sectional study. Our study used a statistical model to predict SFI using the 2009 National Sample Household Survey, where the Brazilian Food Insecurity Scale measured SFI.

Brazil.

The model was applied in a probabilistic sample of 34 003 Brazilians aged 10 years or older that participated in a national dietary survey during 2008–2009. The application of the model generated the probability of each individual being in SFI. The probability of SFI was grouped into quartiles (first quartile with the lowest SFI risk, fourth quartile with highest probability of SFI risk).

The intakes of macro- and micronutrients were associated with SFI. The amount of energy and nutrients in the diet tended to be lower among individuals in the fourth quartile, with highest probability of SFI. The average intake of all studied minerals (Ca, Fe, Na, Mg) was less in individuals in the fourth quartile. Only Na presented a higher percentage of inadequacy in the first quartile, the one with a lower chance of SFI.

The food intake of the Brazilian population at higher SFI risk is characterized by energy reduction, reduced consumption of macronutrients and high prevalence of inadequate micronutrient intakes, as well as a lower mean intakes, when compared with the first quartile with the lowest SFI risk.