Modelos de ponto de mudança em dados de vacinação: aplicação ao número de doses aplicadas em Goiânia, Estado de Goiás e Brasil

Abstract

This study presents the application of changepoint models to analyze the number of vaccine doses administered in Goiânia, the state of Goiás, and Brazil, with the aim of identifying significant changes in time series related to immunization campaigns. Changepoints were detected as significant shifts in the behavior of the data over time, allowing the identification of patterns of increase, decrease, or stabilization in the number of doses applied. For this study, public data were used from the National Immunization Program Information System (SIPNI), under the Secretariat of Health and Environmental Surveillance (SVSA) / Ministry of Health (MS), obtained from the website of the Department of Informatics of the Unified Health System (DATASUS), which is part of the Secretariat for Digital Health and Information (SEIDIGI/MS). The data were extracted using the Tabnet tabulation tool, covering the BCG, Hepatitis B, and Polio vaccines, which are highly representative in terms of basic immunization and are part of the country’s national vaccination schedule. Initially, a preliminary analysis of the data was carried out to understand trends and variations. Then, the results were presented through graphs generated using different statistical packages in R software: changepoint, changepoint.np, bcp, and mcp, the latter two being based on Bayesian models. In addition to the statistical analysis, the identified changepoints were historically contextualized and linked to public health initiatives in Brazil, highlighting their impact on vaccination policies over the years 1994 to 2022. This study is highly relevant as it significantly contributes to the understanding of changes in the number of vaccine doses administered across different regions. It also provides a basis for analysis, enhancement of strategies, and improvement of public policies related to collective health.