Alcohol consumption, wealth, and health - Authors' reply.

Authors’ reply We thank Martin Frisher for his Correspondence about our Article, in which we used linked data to investigate the alcohol harms paradox. We found that increased alcohol consumption was associated with greater harms attributable to alcohol, but that populations of low socioeconomic status (measured by area-based deprivation, income, educational attainment, and social class) were disproportionately affected by alcohol-attributable harm. We found this disparity to be the case even when adjusting for differences in binge drinking, smoking, and bodymass index. We took advantage of the longitudinal nature of our data by excluding people who had a history of alcohol-related and drug-related harm and testing whether our results could be accounted for by social drift (ie, people becoming poor as a result of high-risk alcohol consumption), which we found they could not. Frisher stated that “alcohol use does not necessarily correspond to adverse health outcomes”. Although he referred to results for self-rated general health, our results showed that alcohol consumption was clearly associated with increased risk of alcohol-attributable harms. The most socially advantaged might be at less risk of harm, but consumption across the socioeconomic spectrum still carries a health risk and our findings in no way suggest otherwise. Although we appreciate that other cross-sectional research investigating self-reported outcomes might find less evidence of harm, our study aligns with the wealth of evidence showing the substantial health burdens that arise from alcohol con sumption. Furthermore, the public health message is now far clearer as a result of more causal analyses—in particular, Mendelian randomisation approaches have found no cardiovascular risk protection with low levels of alcohol consumption. Therefore, we feel that the implications for public health are very clear: alcohol is a cause of harm to populations globally and evidencebased actions to address this burden need to be fully implemented. Our study also has implications for the understanding that public health policy makers have of health inequalities. Policies that reduce alcohol consumption across society are likely to disproportionately benefit the poorest in society. It is sometimes suggested that price-based measures (such as taxation or minimum unit pricing) impose greater financial costs on the poorest in society, but our study suggests that their health is likely to benefit the most. Although addressing behavioural risk factors is important to improve population health, this strategy is unlikely to be sufficient to address socioeconomic inequalities in health. There is a continuing need to also focus on addressing underlying social inequalities.