Sistemas de informação em saúde: potencialidades metodológicas no monitoramento da mortalidade neonatal precoce

Abstract

This dissertation consists of four articles and aims at evaluating the use of healthcare information systems as sources for the monitoring of early neonatal death and scientific publication in Brazil. This aim is justified by the importance of the early neonatal component in infant mortality and by the large amount and relevance of information available in the systems. The quantitative analysis is divided in two axes: the authors own use of the databases and the assessment of scientific articles from MEDLINE/PubMEd. Two population based systems were chosen: the Mortality Information System (SIM) and the Live Birth InformationSystem (SINASC); the third system is the record hospital assistance form the SUS network, the Hospital Information System (SIH). The first article describes how the probabilistic record linkage can be used to relate data fron SINASC and SIH in order to asses underreporting of live births on SINASC. The second article compares the results of the linkage between SINASC and SIH with the data made available on the Internet by DATASUS. The third article analyses the multiple causes of earlyneonatal death by means of the Grade of Membership (GoM) method. The fourth article is a bibliometric study articles from MEDLINE/PubMed, and assesses the growth and some characteristics of the scientific publication based on data from SIM, SINASC and SIH in the years of 1990-2006. The results indicate that: 1) the process of making compatible and linkage of database is still considerably difficult; 2) underreporting of live births occurs even in medium and large sizecities in Minas Gerais; 3) the analyses of multiple causes of death by GoM can reveal differences in the profiles of early neonatal deaths according to the hospitals legal statuses; 4) there has been a considerable growth in Brazilian scientific production based on thisinformation systems from the year 2000 onwards. The final comments discuss the main results and highlights possibilities to potencialize the use of the national information systems as sources for scientific research.