> The problem was that "no drinking" group included people that weren't drinking because of poor health They've corrected for this. From the abstract: > Adjus

> The problem was that "no drinking" group included people that weren't drinking because of poor health They've corrected for this. From the abstract: > Adjusting for additional demographic and clinical covariates, and accounting for competing risk of death, did not substantially affect results.