Association between alcohol outlet density and alcohol-related mortality in Wales: an e-cohort study

Abstract

BackgroundLittle is known about the association between alcohol outlet density (AOD) and mortality, and there is a lack of UK longitudinal studies. Results of our previous study on change in AOD and alcohol-related harm to population health (CHALICE) showed that higher availability of alcohol was related to alcohol-related hospital admission. This study aimed to extend this work to investigate the association between AOD and alcohol-related mortality while taking account of migration. MethodsData for the entire population of Wales with a general practice registration on Jan 1, 2006, (n=3 136 881) and follow-up for 24 quarters were record linked using individual identifiers to Office for National Statistics (ONS) mortality files within the Secure Anonymised Information Linkage databank. The Welsh Demographic Service dataset provided information on age, sex, and the lower super output areas (LSOAs) of residence at each of the 24 quarters. To each LSOA at each quarter we linked deprivation fifths using the Welsh Index of Multiple Deprivation 2008 and Rural Urban class as defined by the ONS. AOD was estimated using a mean weighted network distance of all residences to all outlets within a 10 min walk for each LSOA and linked to the dataset by LSOA. Alcohol-related deaths were identified according to the ONS definition. Cox regression models were used to estimate the hazard ratios for mortality associated with quintiles of AOD, adjusting for age, deprivation fifth, and Rural Urban class as time-varying covariates. Migration was accounted for by allowing AOD, deprivation fifth, and Rural Urban class to vary with LSOA moves at every quarter. FindingsOf the 2224 alcohol-related deaths 634 (29%) were in areas of the highest AOD quintile. People living in areas of highest AOD had a higher risk of alcohol-related death than those in the lowest AOD quintile (hazard ratio 1·61, 95% CI 1·37–1·89; p<0·0001). The association between AOD and mortality was non-linear. InterpretationIncreased alcohol availability was associated with a higher risk of alcohol-related death, adjusted for age, sex, and other factors. Although we have included change of outlet density for individuals through quarterly migration between areas, future work should include the change in outlet density in each area over time. FundingPart funded by the National Centre for Population Health and Wellbeing Research and the National Institute for Health Research Public Health Research Programme. The funders had no role in the writing of the abstract or the decision to submit it for publication.