Objective: To develop a social need index for stratification of municipalities and identification of priority areas for reducing fetal mortality.
Methods: ecological study, carried out in the state of Pernambuco, between 2010 and 2017. The technique of factor analysis by main components was used for the elaboration of the social need index. In the spatial analysis, the local empirical Bayesian estimator was applied and Moran's spatial autocorrelation was verified.
Results: The social deprivation index selected two factors that, together, explained 77.63% of the total variance. The preventable fetal mortality rate increased among strata of social need, with rates of 8.0 per thousand births (low deprivation), 8.1 per thousand (medium deprivation), 8.8 per thousand (high deprivation), and 10.7 per thousand (very high social deprivation). Some municipalities in the São Francisco and Sertão Mesoregions had both high fetal and preventable fetal mortality, in addition to a very high social deprivation rate. Conclusion: The spatial analysis identified areas with the highest risk for fetal mortality. The social deprivation index listed some determinants of fetal deaths in areas with worse living conditions. Priority areas for intervention in public policies to reduce fetal mortality and its determinants were detected.