Behavioural and metabolic mediators of socioeconomic inequalities in type 2 diabetes: comparing counterfactual and traditional mediation analysis

There is a well-established social gradient in the occurrence of type 2 diabetes, but the extent to which behavioural or metabolic risk factors explain these inequalities remains unclear. Leveraging data from 7123 adults and over 20 years of follow-up, we used counterfactual mediation analysis to estimate the direct effect of low socioeconomic status (measured as educational attainment and occupational class) on the risk of type 2 diabetes, and the indirect effect through behavioural and metabolic risk factors. Mediators included were smoking, high alcohol consumption, low physical activity, diet low in vegetables or fruits, high body mass index (BMI), high fasting glucose, and hypertension. We compared the results to mediation analysis using the difference and the product of coefficients methods. We found an association between low educational attainment 1.31 (95% CI 1.16, 1.45) and low occupational class 1.24 (95% CI 1.09, 1.38) with future risk of type 2 diabetes. In the counterfactual mediation analysis, behavioural and metabolic risk factors explained 60% (95% CI 41%, 75%) of the effect of low educational attainment and 42% (95% CI 19%, 65%) of the effect of occupational class on the risk of type 2 diabetes. The difference and product of coefficients methods yielded similar results. Well-established behavioural and metabolic mediators explained roughly half of the health inequalities in the incidence of type 2 diabetes. Public health interventions should consider alternative mechanisms to reduce disparities in the incidence of type 2 diabetes.