Analysis of the application of a deterministic routine for identifying multiple pregnancies on Live Birth Information Systen, Brazil

Abstract

Objective: to evaluate the application of a deterministic routine for identifying multiple pregnancies on the Brazilian Live Birth Information System (SINASC). Methods: SINASC data deduplication and linkage with the mortality database (fetal deaths) for Rio de Janeiro state for the period 2007-2008; we used a deterministic routine, using a key based on SINASC maternal and birth information, complemented by manual review. Results: of the 433,874 SINASC records, 9,036 (2.1%) were classified as multiple pregnancy newborns; after implementing the routine, we reclassified 385 records as twins, and 286 as singletons; accuracy of multiple pregnancy information on the SINASC database was high (sensitivity= 95.8%; specificity=99.9%); applying the routine without the manual review process increased sensitivity by 4.2%, with no significant change of specificity. Conclusion: despite the accuracy of information regarding multiple pregnancy held on SINASC, we suggest the use of this routine as an option for improving classification of twins.