This study aimed to investigate the performance of the Hospital Information System of the Brazilian Unified National Health System (SIH-SUS) in identifying cases of maternal near miss in a hospital in Rio de Janeiro, Brazil, in 2008. Cases were identified by reviewing medical records of pregnant and postpartum women admitted to the hospital. The search for potential near miss events in the SIH-SUS database relied on a list of procedures and codes from the International Classification of Diseases, 10th revision (ICD-10) that were consistent with this diagnosis. The patient chart review identified 27 cases, while 70 potential occurrences of near miss were detected in the SIH-SUS database. However, only 5 of 70 were "true cases" of near miss according to the chart review, which corresponds to a sensitivity of 18.5% (95%CI: 6.3-38.1), specificity of 94.3% (95%CI: 92.8-95.6), area under the ROC of 0.56 (95%CI: 0.48-0.63), and positive predictive value of 10.1% (IC95%: 4.7-20.3). These findings suggest that SIH-SUS does not appear appropriate for monitoring maternal near miss.