Beneficial associations between low intensity alcohol consumption and all cause mortality may in part be attributable to inappropriate selection of a referent group and weak adjustment for confounders. Data from up to 10 cohorts representative of the English population, suggest [...]