Background Socio-economic inequalities exist in prevalence and incidence of type-two diabetes (T2D). Hypothesised explanations in the literature refer to differences on the basis of socio-economic position (SEP) in experience of material deprivation, psychosocial stress and health behaviours. We test the hypothesis that each of these pathways mediated the association between SEP and T2D, and vary on the basis of area.

Methods Using the British Household Panel Survey (BHPS), a longitudinal study which begun in 1991, we looked at factors influencing self-reported T2D incidence. The sample was restricted to participants satisfying the following criteria: a) British residents of white ethnicity; b) non-diabetic; c) aged 30–70 on entry to the survey; and d) present in waves 10–18 of the survey. These criteria result in an analytical sample of 6,641 participants. The association between T2D and SEP is decomposed into direct and indirect effects using the Karlson, Holm and Breen (KHB) decomposition technique. Models with random effects for area were run separately for England, Scotland and Wales – due to differences in deprivation scores – using Stata 14.

Results Evidence was found for an association between SEP and T2D. The total effect (coefficient (β) 0.320, standard error (SE) 0.171) was decomposed into direct and indirect effects through the material, psychosocial and health behaviour pathways. The contribution of these mediators to the total effect was significant and explained approximately one third of the total effect. This was mostly explained by health behaviours; exercise (β 0.05, SE 0.02) and obesity (β 0.04, SE 0.02). The material and psychosocial pathways did not contribute significantly. Random effects for area did not improve the model fit for each region (p > 0.05).

Conclusion These findings show the association between SEP and T2D is mediated by exercise and obesity, confirming the findings of previous studies. This study was limited by imperfect measures of SEP and each pathway. Policy interventions aimed at promoting healthy lifestyle could focus on making this accessible to low SEP groups in order to reduce inequalities in T2D. Further research in this area would benefit from measuring a greater array of health behaviour.