Background Multimorbidity, defined as experiencing two or more long-term health conditions, is a common and increasing challenge for global health systems. Multimorbidity is associated with socioeconomic disadvantage; however, the extent to which common risk factors explain these inequalities is unclear. We examined the potential contribution of five behaviour-related risk factors in explaining socioeconomic inequalities in developing multimorbidity across the life course.
Methods The Twenty-07 cohort study recruited participants, aged 15, 35 and 55 years at baseline in 1987 (total n = 4,510), who were representative of the west of Scotland population. Participants were followed up over twenty years, with data collected at five time points using nurse-led interviews. Our primary outcome was current multimorbidity, with five risk factors (smoking, alcohol consumption, diet, physical activity, body mass index (BMI)) and two measures of socioeconomic status (area-based deprivation and household income) assessed as predictor variables. Multilevel logistic regression models (with observations nested within individuals, within postcode sectors) and predicted probabilities were estimated. The magnitude of socioeconomic inequalities in multimorbidity was calculated using relative indices of inequality, and attenuation through the addition of risk factors was assessed. Multiple imputation with chained equations was conducted. All analyses were conducted in Stata v13.
Results As expected, socioeconomic disadvantage was associated with increased risk of developing multimorbidity (e.g. most versus least deprived areas: OR 1.48, 95% CI: 1.23, 1.77) throughout the life course. Smoking, diet and BMI were strong independent predictors of developing multimorbidity (e.g. relative to healthy BMI, OR 1.88, 95% CI: 1.42–2.49 for ‘morbidly obese’ and OR 1.41, 95% CI: 1.19–1.67 for ‘obese’). A dose-response relationship was also observed with the number of risk factors. Adjusting for the five risk factors only explained 41% of the socioeconomic inequalities in developing multimorbidity; a substantial socioeconomic gradient remained. Sensitivity analyses showed similar results for household income, different definitions of multimorbidity, and a complete case analysis.
Discussion Socioeconomic inequalities in the development of multimorbidity exist even after taking account of known risk factors. Our study longitudinally investigates multiple risk factors across the adult lifecourse; nevertheless important limitations, common to many cohort studies, include self-reported data and attrition. Addressing these risk factors, particularly obesity and smoking, could reduce the future burden of multimorbidity and help address inequalities. However, preventive measures are unlikely to fully address the greater burden of multimorbidity amongst socioeconomically disadvantaged populations therefore a healthcare system responsive to this is essential.