OBJECTIVE: The present contribution aims at illustrating the methodological problems that may arise from investigating the indirect social costs of alcohol consumption. METHODS: In order to estimate the alcohol-related productivity losses through excessive morbidity, regression analysis was performed, using data from the 2(nd) Swiss Health Survey (1997). For the purpose of sensitivity analysis two approaches were used (a) Rice and b) Harwood), and all models were calculated with and without smearing-adjustment as well as with and without adjustment for confounding variables. RESULTS: The present study shows that small changes in method may result in major variations in the estimated total costs. In monetary terms, the indirect costs for alcohol consumption in Switzerland through excess-morbidity were estimated at between 0.119 billion Francs (model Harwood, smearing-adjusted) and 2.175 billions Francs (model Rice, adjusted for confounding variables). CONCLUSION: Since many methodological problems in estimating the indirect costs of alcohol consumption have not been adequately resolved yet, results of those studies have to be interpreted with caution.