Here's an example of the "moderate drinking is healthy" risk curves that the article refers to: https://publications.parliament.uk/pa/cm201012/cmselect/cmsctech

Here's an example of the "moderate drinking is healthy" risk curves that the article refers to: https://publications.parliament.uk/pa/cm201012/cmselect/cmsctech/1536/1536vw08.htm https://publications.parliament.uk/pa/cm201012/cmselect/cmsc... This newer work seems basically to be arguing that the effect on the lefthand size of the curve is because drinking alcohol is so common that people who drink nothing are likely to be in poor health already.