Spatial and Spatio-temporal Epidemiology · Oct 15, 2026

Bayesian spatio-temporal analysis of all-cause mortality across neighborhood deprivation measures in Sweden

By van der Velde L, Baravelli CM, Mattsson M et al.

Despite declining mortality rates in Sweden, significant geographic and socioeconomic disparities persist. Previous research has typically employed single socioeconomic indicators, which may not adequately capture the multidimensional nature of deprivation. To address this limitation, this study applied the Index of Multiple Deprivation in Sweden (IMDIS). Using a granular approach based on Demographic Statistical Areas (DeSO), we examined how small-area deprivation, measured through the overall composite IMDIS score and its underlying domains and indicators, relates to all-cause mortality by sex. We conducted an ecological cross-sectional study using Swedish register data between 2015 and 2019, totalling 50,400,585 person-years across 5983 DeSO-areas, including the total resident population across all age groups. We applied Bayesian spatio-temporal models to assess associations between all-cause mortality and IMDIS scores, domains (Education, Income & Capital, Employment, Housing), and 15 individual indicators of deprivation using Mortality Rate Ratios (MRRs). The Relative Index of Inequality (RII) was calculated to quantify the magnitude and direction of the deprivation gradient. A consistent deprivation gradient was observed where, for the overall IMDIS score, the most deprived areas experienced a 43% higher mortality risk (MRR 1.43; 95% CrI: 1.39-1.46) compared to the least deprived. Relative inequalities were more pronounced in men (1.51, 1.47-1.55) than in women (1.34, 1.29-1.38). RII values indicated systematic gradients across nearly all indicators, except for preschool attendance in women, with the highest inequality found when selecting disposable income as a measure of deprivation for men. All-cause mortality in Sweden follows a clear deprivation gradient in both men and women, with higher risks in the most deprived areas. The associations also demonstrate that the composite IMDIS measure is not merely the sum of its individual domains or indicators, but reflects a broader, underlying deprivation construct that captures their combined pattern of association.

doi.org/10.1016/j.sste.2026.100853Open accessPubMed 43051864Populations

AgingResearch.News