Cobertura vacinal, preferência política e demografia: um estudo ecológico dos municípios nas eleições de 2022

Abstract

Political preference might be one of the determinants of vaccine coverage, and this monograph aims to test this hypothesis of association when considering the effects of other demographic variables on the vaccine coverage of 5 pediatric vaccines in the Brazilian municipalities. Methodology: The present monograph consists of an epidemiological study with ecological design that evaluates the relationship between the proportion of the electorate that voted for the candidates Bolsonaro and Lula during the Brazilian presidential elections of 2022 and the vaccine coverage for 5 pediatric vaccines in that same year using linear regression models with a compositional approach. The sample and level of analysis consisted of the 5570 Brazilian municipalities, and the models were adjusted for sociodemographic variables of interest. Theoretical Background: In Section 2.2, a brief description of the properties of ecological studies and their limitations, principles of linear regression and introductory concepts in compositional analysis is provided. In chapter 3, the vaccines recommended by the National Immunization Program are listed, the concept of vaccine coverage is defined and an overview of the evidence in academic literature on possible determinants of vaccination rates and vaccine hesitancy in Brazil and abroad is presented, including both ecological studies and texts whose designs include individual-level samples. Results: A statistically significant, negative association was found between the share of Lula voters in a municipality and the coverage for 4 vaccines, along with a significant association of the same sign for the share of Bolsonaro voters for 2 vaccines. In both cases, the effect sizes are small, and lower than the ones produced by the control variables, such as municipalities’ population, age and educational structure. Other findings include: less than half municipalities reached vaccination rate targets; a large share of vaccine coverage higher than 100% indicating a possible problem in the calculation of the denominator; similar results in alternative models; the inability of the sociodemographic variables (including electoral adherence) to sufficiently explain the variation of vaccine coverage between the Brazilian municipalities. Conclusion: Political preference does not appear to be an important determinant of coverage for the vaccines under analysis when considering Brazilian municipalities as the sample unit. Future research is needed to identify models with higher predictive accuracy and possible occult variables that are determinants of vaccine coverage.