Neighbourhood built environments, socioeconomic position, and hospital admissions for cardiovascular disease: a prospective study using UK Biobank

BackgroundNeighbourhood environments may influence cardiovascular disease (CVD) risk, e.g. by influencing diet and physical activity (PA) behaviours. We explored whether associations between characteristics of neighbourhood environments and CVD are modified by area deprivation and household income. If effects of neighbourhood risk exposures vary by socioeconomic position, efforts to improve population health by improving neighbourhood built environments could widen health inequalities. MethodsIn the UK Biobank cohort we used linked records of hospital admissions to assess the relative hazard of being admitted to hospital with a primary diagnosis of CVD according to three characteristics of the neighbourhood built environment: availability of formal PA facilities, proximity of a takeaway/fast-food store, and neighbourhood greenspace. We then examined potential effect modification of the main associations by household income and area deprivation. We used Cox proportional hazards models, adjusted for likely confounding, and calculated relative excess risks due to interaction (RERI) to assess effect modification on the additive scale. We also examined the combined modifying role of income and deprivation. ResultsThere were 13,809 incident CVD admissions in the sample (mean follow-up=6.8 years). Overall associations between neighbourhood exposures and CVD-related hospital admissions were weak to null. However, there was evidence of effect modification by both area deprivation and household income. Greater availability of PA facilities near home was associated with lower risk of CVD-related admission in more deprived areas, but only among people in higher-income households. Area deprivation and household income both modified the association with fast-food proximity. More greenspace was not associated with lower risk of CVD-related admission for any group. Some results differed between women and men. Findings were largely robust to alternative model specifications. ConclusionsImproving deprived neighbourhoods by increasing the number of PA facilities, while also ensuring access to these is free or affordable, may improve population health. Examining effect modification by multiple socioeconomic indicators in parallel can yield deeper insight into how different aspects of the peoples socioeconomic conditions influence their relationship with the built environment and its effects on their health. Improved understanding may help to avoid generating or perpetuating health inequalities when neighbourhood-based built environment interventions are designed.