Mineração de dados em saúde: perspectivas para a vigilância em saúde da população exposta ao agrotóxico

Abstract

The extensive use of pesticides in Brazilian agriculture culminates in a greater exposure of the population to these substances to a greater or lesser extent, through different routes and routes of exposure. Therefore, the Ministry of Health, in 2012, with Ordinance GM/MS No. 2,938, implemented the Surveillance of populations exposed to pesticides (VSPEA). However, the detection of chronic intoxication and the link with other diseases from health information systems is a challenge for the Health Surveillance system in its basic principles of prevention, control and health promotion. The general objective of the research is to discuss the factors associated with pesticide poisoning based on data mining applied to databases of health information systems. From the application of statistical treatments and data mining to SINAN, SIM and SINASC databases to extract important associations that could explain the impacts of pesticide poisoning. The research findings consist of data collection to understand the context resulting in the construction of a data-wharehouse focused on decision-making and planning of the VSPEA, in the modeling of datasets from health information systems, enabling the predictive modeling of three prediction matrices capable of recognizing patterns of associated factors for the type of exposure, classification of poisoning and evolution of pesticide poisoning. It reveals the potential of interdisciplinarity with Data Science for health surveillance, as well as raising questions for future research and reflections on the process of epidemiological surveillance on the subject.