Mineração de dados com regras de associação aplicada em dados de unidade de saúde de pronto atendimento

Abstract

This work aims to apply the steps of the process of the discovery of knowledge in an electronic patient record from a public health unit. The discovery of correlated diseases in medical records can be difficult or time-consuming for the healthcare professional if the amount of data is immense or when there is no availability of specialized computer systems. Data mining techniques, such as rules of associations, clustering, or classifications, are alternatives for obtaining a specialized system that allows interpreting data using patient groups. To do this, we used a set of data extracted from a database of an electronic medical record of single health system users, containing 43,879 patients and 2,296,626 patients. Experimental results show that the association of diseases and groups of patients can help health professionals and managers in the application of prevention policies, which can improve the quality of life of a region as well as the economy in the treatment of new cases.