Identificação de pontos de corte otimizados para a classificação da insegurança alimentar domiciliar pela Escala Brasileira de Insegurança Alimentar em diferentes grupos populacionais

• Ano de defesa: 2017
• Autor(a) principal: Interlenghi, Gabriela dos Santos
• Orientador(a): Reichenheim, Michael Eduardo
• Banca de defesa: Segall-corrêa, Ana Maria , Hökerberg, Yara Hahr Marques , Moraes, Claudia Leite , Sichieri, Rosely
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade do Estado do Rio de Janeiro
• Programa de Pós-Graduação: Programa de Pós-Graduação em Saúde Coletiva
• Departamento: Centro Biomédico::Instituto de Medicina Social
• País: BR
• Palavras-chave em Português: Insegurança alimentar; Escalas; Psicometria; Inquéritos e Questionários; Modelos estatísticos; Análise multivariada
• Palavras-chave em Inglês: Food insecurity; Scales; Psychometrics; Surveys and Questionnaires; Statistical models; Multivariate Analysis
• Área do conhecimento CNPq: CNPQ::CIENCIAS DA SAUDE::SAUDE COLETIVA
• Link de acesso: http://www.bdtd.uerj.br/handle/1/4771

The main objective of this Thesis was to examine the classification of household food insecurity (FI) by the Brazilian Household Food Insecurity Measurement Scale (EBIA) with respect to its ability to separate homogeneous groups corresponding to increasing ordinal severity levels of FI across different population domains. Additionally, this Thesis investigated the feasibility of appropriately classifying families with children and/or adolescents from the 8 general/adult items only. Data derived from two surveys. The first concerns a cross-sectional population-based study, conducted in 2010, with a representative sample of 1105 households from the district of Campos Elíseos, municipality of Duque de Caxias, state of Rio de Janeiro, Brazil (Pesquisa SANDUC 2010). The second comprises the National Household Sample Survey of 2013 (PNAD-2013), which is a nationally representative survey with a final sample of 116,543 household units. To meet the main purpose, the 14 dichotomic items of the current version of the EBIA were assessed, both at Pesquisa SANDUC 2010 (Article 1) and PNAD-2013 (Article 2). The additional objective considered the analysis of the 8 general/adult items with data from PNAD-2013 (Manuscript). Latent Class Factor Analysis (LCFA) models were applied to the answers to EBIA s items to identify cut-off points able to accurately classify latent groups corresponding to different levels of FI. Results of the first article indicated four homogeneous groups of FI, with a very high degree of class separation (entropy= 0.906), whose model-based cut-off points broadly endorse those traditionally employed. However, whereas families with only one affirmative answer are usually classified as mild FI, LCFA showed that these families would be better classified if placed in the same stratum than those without any affirmative response. The second article corroborates findings of the former regarding the classification of FI. Accordingly, the following optimal cut-off points were identified in the aggregate Brazil: 1/2, 5/6 and 10/11 in households with children and/or adolescents (raw score= 0-14 points); 1/2, 4/5 and 6/7 in households with only adults (raw score= 0-8 points). From the analysis of the 8 general/adult items only, the results of the third study showed the identification of a single set of cut-off points across all population domains, regardless of their demographic composition. These cut-off points are the same as those identified for households with only adults in the second paper. Moreover, these findings clearly show the possibility to capture the severity of FI in households with children and/or adolescents even without using their specific items. It is expected that evidence from this Thesis contributes to discussions on the most suitable cut-off points for the EBIA and the possibility of simplifying the measurement of FI by the instrument in certain contexts, aiming the appropriate classification of families and, consequently, a correct assessment and monitoring of FI.