
There are well-known problems inherent in most observational studies of alcohol and health. As well as having to rely on self-reporting of intake, these include single-time point consumption assessments, inadequate confounder adjustments, accounting for reverse causation, selection bias, short follow-up, residual confounding and lacking information on chronic health conditions that may mediate the association between alcohol consumption and mortality.