The effects of regional characteristics on alcohol-related mortality -: a register-based multilevel analysis of 1.1 million men

The aim of this study is to assess to what extent selected characteristics of functional regions affect alcohol-related mortality among men in Finland after adjusting for individual-level characteristics. The study was conducted as a multilevel Poisson regression analysis, with individuals (n = 1.1 million) as the first level and functional regions of Finland (n = 84) as the second level. The analysis covered men aged 25-64. The data are based on the 1990 census records, which were linked to death records in 1991-1996. The outcome measure was alcohol-related mortality, which was defined using information on the underlying and contributory causes of death. The individual-level covariates included age, education, socioeconomic status, marital status and mother tongue. The area-level variables considered were the proportion of manual workers, unemployment level, median household income, Gini coefficient of income, family cohesion, voting turnout, level of urbanisation and proportion of Swedish-speaking inhabitants. A high proportion of manual workers and of unemployed and low social cohesion (family cohesion and voting turnout) were found to produce adverse effects on alcohol-related mortality, and the independent effects of these variables remained after adjustment for all individual-level and area-level characteristics. The protective effect of high level of urbanisation was revealed after adjustment for other individual- and area-level characteristics. Neither mean income nor income inequality were related to alcohol-related mortality. Adjusting for individual-level variables diminished the average relative deviation of alcohol-related mortality among the functional regions by 41%. The inclusion of area-level characteristics in the model resulted in a total diminution of variation of 79%. The area characteristics considered in this study had a notable effect on alcohol-related mortality, although these effects were smaller than those of the individual-level characteristics. Fuller understanding of the mechanisms underlying the effects of area measures of social structure and cohesion on risky alcohol consumption and alcohol-related mortality is needed. (C) 2003 Elsevier Ltd. All rights reserved.