Onset of Type 2 diabetes in adults aged 50 and older in Europe: an intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy.

Abstract

Background: Disparities in the development of Type 2 Diabetes (T2D) are associated with various social determinants, including sex/gender, migration background, living arrangement, education, and household income. This study applied an intersectional perspective to map social disparities and investigate intersectional effects regarding the onset of T2D among older adults across Europe.

Methods: We used data from the Survey of Health and Retirement in Europe (SHARE) to conduct an Intersectional Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (I-MAIHDA) of T2D onset. Individuals aged 50 years or older without known T2D at Wave 4 (2011, baseline) were included and followed through Waves 5 (2013), 6 (2015), 7 (2016), and 8 (2019-2020). Intersectional models were used to estimate additive main effects of sex/gender, migration background, living arrangement, education level, and household income and intersectional interactions.

Results: A total of 39,108 individuals were included (age at baseline M = 65.18 years (SD = 9.62), 57.4% women). T2D onset was reported for 9.2% of the sample over the 9-year observation period. In the fully adjusted model, all social determinants showed significant additive associations with T2D onset, while the discriminatory accuracy of the social strata was found to be low (Variance Partition Coefficient = 0.3%).

Conclusions: This study provides a comprehensive mapping of intersectional disparities in onset of T2D among older adults in Europe. The results highlight the risk heterogeneity within the population and show social disadvantages faced by certain groups. However, while the T2D risks were higher in some strata than in others, the intersectional effects were small overall and mostly attributable to the additive main effects. The results suggest that public health strategies to prevent T2D should be universal but tailored to meet the specific situation of the different intersectional strata.