TECHNOLOGY IN HEALTH: KNOWLEDGE DISCOVERY IN PUBLIC HEALTH DATABASES: STUDY OF VIRAL HEPATITIS IN THE STATE OF PARANÁ, BRAZIL

Autores

  • Carla Machado da Trindade Pontifical Catholic University of Paraná
  • Claudia M. Cabral Moro Pontifical Catholic University of Paraná
  • Márcia Gil Aldenucci Paraná State Health Department
  • Júlio César Nievola Pontifical Catholic University of Paraná
  • Deborah Ribeiro Carvalho Pontifical Catholic University of Paraná
  • Samuel Jorge Moysés Pontifical Catholic University of Paraná

Palavras-chave:

Public Health Information Systems, Knowledge Discovery in Databases, Data Mining, Epidemiological Surveillance, Viral Hepatitis

Resumo

This paper shows some benefits that the methodology of Knowledge Discovery in Databases (KDD), using the data mining technique, can bring when used in databases that store data on the health of population, describing the analytical process applied in the identification of the behavior of viral hepatitis. The KDD process involved the application of the classification technique on 2003 data, stored in the Notifiable Diseases Information System of Paraná Health Department. Sixty-five characteristics of 3.063 investigation forms were analyzed, resulting in 4 decision trees and 99 classification rules. Of these rules, 60 were analyzed and the other ones were discarded because they did not contemplate enough examples to be considered valid or they had a high number of errors. The method enabled the database to be explored thoroughly, as well as enabling an increased number of appraised characteristics, the identification of problems relating to the quality of the data and to information routinely used by the epidemiological surveillance service. It was also possible to discover the occurrence of hepatitis B in its chronic form in children under 13 years old. This knowledge, unperceived in the original database, can help in the formulation of new policies, suggesting the importance of this method as an form of scaling up routine strategies, with the aim of reducing and controlling diseases. This technology contributes towards the exploration of the data stored by the various Health Information Systems.

Downloads

Edição

Seção

Papers