Reprodutibilidade de um instrumento simplificado para triagem de insegurança alimentar em famílias brasileiras

Abstract

Food insecurity (FI) has affected an increasing number of Brazilian families in recent years, being characterized by the lack of regular and permanent access to quality food in sufficient quantity. The consequences resulting from this condition involve worsening nutritional, mental, and health status as a whole. The Brazilian Food Insecurity Scale (EBIA) is an important FI measurement tool, providing information on the perception/experience of hunger in a Brazilian household. The EBIA is routine data collection from large-scale national surveys and regional and local surveys. However, due to the nature and purpose of these surveys, it is observed that the individual identification of families who need immediate help is impaired. Thus, the use of a brief FI screening instrument based on the EBIA would speed up the identification of these families, through their incorporation into the routine of clinical and care activities. With this in mind, Brazilian researchers developed the Screening for Risk of Food Insecurity (TRIA), a 2-item instrument to screen families with children under 5 years old at risk of FI. Additional studies are needed to assess the applicability of TRIA in households with other settings. In this context, this dissertation was prepared with the purpose of contributing to the understanding of FI, as well as providing subsidies for its evaluation, monitoring and confrontation in Brazil. To this end, a literature review was carried out, addressing aspects of FI, and an original article was produced. The latter aimed to test the reproducibility of the TRIA, analyzing its performance in different strata of the Brazilian population. For this, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of the 2-item TRIA compared to the gold standard, the 14-item full EBIA were calculated using data from the National Survey by Household Sample (PNAD-2013). In addition, a graph was generated with the PPV adjusted by the FI prevalence for each Brazilian state. The convergent validity of the instrument tested the strength of the association between FI and dietary variables. The combination of questions 2 and 4 showed the best performance in all population strata analyzed, providing sensitivity ≥86%, specificity ≥87%, accuracy ≥88%, PPV ≥62% and NPV ≥97%, confirming reproducibility of the TRIA. The adjusted PPV ranged from 42.8% to 87.6%, increasing with the higher prevalence of FI by state. The TRIA showed high agreement when compared to the EBIA in all strata of the Brazilian population. Thus, the TRIA is recommended to identify families at risk for FI, especially in high-risk population subgroups, and its use is recommended in clinical practice, in emergency or public calamity situations, and in health surveillance in Brazil, providing a basis for planning, managing and strengthening programs and public policies aimed at FI.