A matter of measurement? A Swedish register-based study of migrant residential segregation and all-cause mortality.

Background: In recent decades, Sweden has become an increasingly diverse society by origin, but one in which residential segregation by migrant background has also increased. This study examines how different aspects of migrant residential segregation are associated with all-cause mortality among native-born and migrant populations. Methods: Using Swedish population-based registers, this longitudinal open cohort study assessed associations between four local level indices of migrant residential segregation and all-cause mortality among adult migrant and native-born residents of Sweden's three largest metropolitan areas (Stockholm, Gothenburg, and Malmö) between 2004 and 2016. Multilevel Poisson regression models, adjusted for individual-level sociodemographic and socioeconomic factors as well as area-level socioeconomic conditions, were used to estimate associations between these indices and all-cause mortality. Results: Moderate decreased mortality risks were observed among migrants in residential areas with higher levels of migrant density, isolation, and exposure in fully adjusted models. However, isolation and exposure effects could not be distinguished due to a high degree of correlation between the isolation and exposure measures. In fully adjusted models mortality gradients were largely unobserved among native-born individuals in relation to migrant residential segregation. The evenness dimension of segregation showed limited relevance for mortality risk in both groups. Conclusions: This study provides evidence that higher migrant density is associated with lower mortality risks among migrants, suggesting that residential areas with higher proportions of migrants may offer health benefits for migrants. These findings highlight the importance of residential contexts in shaping migrant health outcomes.