Estimating the Real Burden of Gestational Syphilis in Brazil, 2007 to 2018: Underreporting Correction in Official Data

Background: Although several studies have estimated gestational syphilis (GS) incidence in several countries, underreporting correction is rarely considered. This study aimed to estimate the level of under-registration and correct the GS incidence rates in the 557 Brazilian microregions, being the first study on quality of GS data at a national level. Methods: Brazilian GS notifications between 2007 and 2018 were obtained from SINAN-Syphilis system. A cluster analysis was performed to group microregions according to quality of GS notification. A Bayesian hierarchical Poisson regression model was applied to estimate the reporting probabilities among the clusters and to correct the associated incidence rates. Findings: We estimate that 45 196 (90\%-HPD: 13 299; 79 310) GS cases were underreported in Brazil from 2007 to 2018, representing a coverage of 87·12\% (90\%-HPD: 79· 40\%; 95·83\%) of registered cases. Underreporting level differs across the country, with microregions in North and Northeast regions presenting the highest percentage of missed cases. After underreporting correction, estimated GS incidence rate in Brazil increased from 8·74 to 10·02 per 1000 live births in the period. Interpretation: Our findings highlight disparities in registration level and incidence rate of GS in Brazil, reflecting regional heterogeneity in the quality of syphilis surveillance, access to prenatal care and childbirth assistance services. This study provides robust evidence to enhance national surveillance systems and guide specific policies for GS detection, disease control, and potentially mitigate the harmful consequences of mother-to-child transmission. The applied methodology might be applied in other regions to correct disease underreporting. Funding: The Syphilis Project receives funding from CNPq/Brazil (Grant number 442873/2019-0). CIDACS/Fiocruz-Bahia is supported by grants from CNPq/MS/The Gates Foundation (401739/2015-5) and the Wellcome Trust (202912/Z/16/Z). EPS is funded by the Wellcome Trust (213589/Z/18/Z). The funder of the study had no role in study design, data collection, management, analysis, interpretation, or writing of the report.Declaration of Interests: We declare no competing interests.Ethics Approval Statement: Not applicable. All analyses used de-identified and publicly available data.