LAUSR

INTRODUCTION AND AIMS Unrecorded alcohol consumption has increased strongly in Finland after 1995 when the country joined the European Union. This development may have rendered alcohol sales less trustworthy as a proxy for population drinking, and less powerful as predictor of alcohol-related harm. The study aims to test this contention by analyzing the association between recorded and unrecorded alcohol consumption on the one hand, and alcohol-specific mortality on the other. DESIGN AND METHODS We analysed age-standardised rates of alcohol-specific deaths for the working-age (15-64 years) population. For alcohol consumption, we used (i) alcohol sales in litres of 100% alcohol per capita, and (ii) estimated unrecorded consumption in litres of 100% alcohol per capita. The data spanned the period 1975-2015. As the data were cointegrated, the relations between mortality and the alcohol indicators were estimated through time-series analysis of the raw data. RESULTS A one litre increase in alcohol sales was associated with an increase in alcohol-specific deaths of 7.590 deaths per 100 000; the corresponding figure for unrecorded consumption was 9.112 deaths per 100 000. Both estimates were statistically significant (P < 0.001), but the difference between them was not significant (P = 0.293). Although recoded consumption captured the main feature of the trends in alcohol-specific mortality, it accounted for only half of its marked increase in 1975-2007, while unrecorded consumption explained the remaining part. DISCUSSION AND CONCLUSIONS Our study confirms previous findings that recorded alcohol consumption is an important determinant of alcohol-specific mortality in Finland. A more novel insight is the importance of unrecorded consumption in this context.