Núcleo de Indicadores Sociais, Fundação de Economia e Estatística. R. Duque de Caxias 1.691, Centro Histórico. 90010-283 Porto Alegre RS Brasil. [Email]
This study assesses the quality of the SIM and SINASC information systems in coverage, incompleteness and consistency aspects, as well as the contribution of the linkage for data retrieval. It includes all live births and infant deaths in Rio Grande do Sul from 2000 to 2014. The records were paired by deterministic linkage through the DNV number and, in its absence, by probabilistic linkage. SINASC's coverage rose from 72.2% in 2000 to 98.9%, namely a 37% increase in the number of matched records. All variables in SINASC presented excellent incompleteness throughout the period, except for the number of dead children and maternal occupation. SIM presented poor or very poor incompleteness for most of the variables until 2003. Although it improved, in 2014, six variables still presented regular or poor incompleteness. The linkage procedure greatly reduced the incompleteness for most variables. There was a great variability in terms of consistency: while for gender this percentage was over 97% throughout the period, for another five variables it was still less than 75% in 2014. SINASC presented high coverage level and excellent incompleteness. Problems related to consistency persist. This study shows the linkage technic efficiency to retrieve missing information.