Background Diabetes prevalence is high worldwide, affecting entire populations. Yet some population groups are more susceptible than others. In contemporary western societies, socioeconomically disadvantaged groups are disproportionally affected. The pattern is less clear for diabetes mortality. Researchers argue increasingly in favour of looking beyond individual socioeconomic position (SEP). In light of the above, this paper looks at the relationship between individual-based and household-based SEP measures and diabetes mortality.

Methods Data consisted of the Belgian 2001 census for the Flanders region linked to register data on cause-specific mortality during the period 2001–2010. The study population included all 35-year-old to 74-year-old official inhabitants of Flanders married or cohabiting at the 2001 census. Age-standardised death rates (direct standardisation) and mortality rate ratios (MRRs; Poisson regression) were computed for the different SEP groups.

Results Inverse gradients were observed by own education, partner's education and housing status (combination of housing tenure and quality). For example, among 55-year-olds to 74-year-olds, MRRs according to partner's education were 3.0 (95% CI 2.4 to 3.8) for women with a lower-educated partner and 1.6 (95% CI 1.2 to 2.0) for women with a higher secondary-educated partner, relative to women with a higher-educated partner. As for housing status, diabetes-related mortality was particularly high among tenants and low-quality owners. The association between each of the SEP measures and diabetes-related mortality remained after adjusting for the other measures.

Conclusions There were large differences in diabetes-related mortality according to both individual-based and household-based SEP measures, indicating the importance of the individual and household levels for understanding socioeconomic inequalities in diabetes mortality.