Maternal mortality risk among Black women in Brazil: a nationwide spatial consensus approach

Although preventable, maternal mortality remains a global challenge. Its burden is disproportionately concentrated in low- and middle-income settings. Substantial subnational geographical variations also persist within countries, driven by underlying socioeconomic disparities and unequal access to quality healthcare services. Here we propose a spatial consensus framework combining frequentist and Bayesian models to overcome variance instability in small-area estimates, mapping maternal mortality risk among Black women in Brazil to guide equitable public health interventions. Using nationwide surveillance data (2019 - 2024), we applied a spatial Poisson probability model to detect high-rate clusters of maternal mortality among Black women. Concurrently, Bayesian hierarchical spatial models were fitted to smooth small-area rates and stabilise variance. Results from both approaches were integrated into a five-level risk typology to categorize Brazilian municipalities. National aggregate reductions in Brazilian maternal mortality mask profound, spatially concentrated racial inequities. The spatial classification of risk scenarios identified 13 municipalities with Confirmed Critical Risk, 28 municipalities with Persistent Regional Risk, 12 municipalities with a Local Excess Alert, and 4 exhibiting Epidemiological Silence. The vast majority remained at the Expected Risk Level. The proposed spatial consensus framework translates complex spatio-temporal modelling into an operational typology for surveillance systems. This approach bridges the gap between spatial analytics and public health decision-making, providing a replicable tool to address structural racial disparities in maternal health within decentralised health systems.