Additional file 2 of The effects of exceeding low-risk drinking thresholds on self-rated health and all-cause mortality in older adults: the Tromsø study 1994–2020

Additional file 2: Table S2. All-cause mortality risk by alcohol consumption according to cohort. The Tromsø Study 1994–2020. HR; hazard ratios, are based on cox proportional hazard models with repeated measures of alcohol consumption. All time-varying scores were updated in 2001, 2007–08, and 2015–16 for those who participated. All estimates are adjusted for education, age and including all listed covariates. End of follow-up on November 25, 2020. Exponentiated coefficients; 95% confidence intervals in brackets. *p <  0.05, **p <  0.01, ***p <  0.001. aIn 1994–95, the seven-item CONOR Mental Health Index (CONOR-MHI) was used, whereas in the three subsequent surveys, the ten-item Hopkins Symptom Check List-10 (HSCL-10) was used bHII measures somatic diseases according to the impact that each condition has on SRH. cSubjects reporting the use of either or both sleeping pills/tranquilisers. In 1994–95, the time frame asked was “during the last two weeks”, while in the three subsequent surveys it was “during the last four weeks”. cSubjects reporting the use of either or both sleeping pills/tranquilisers. In 1994–95, the time frame asked was “during the last 2 weeks”, while in the three subsequent surveys it was “during the last 4 weeks”.