Citation

Hemström O. Eur. J. Public Health 2002; 12(4): 254-262.

Affiliation

Centre for Health Equity Studies (CHESS), Stockholm, University/Karolinska Institute, SE-106 91 Stockholm, Sweden. orjan.hemstrom@chess.su.se

Copyright

(Copyright © 2002, Oxford University Press)

DOI

unavailable

PMID

12506500

Abstract

BACKGROUND: This study aims at estimating the contribution of alcohol to socioeconomic mortality differentials in Sweden. METHODS: Data were obtained from a Census-linked Deaths Registry. Participants in the 1980 and 1990 censuses were included with a follow-up of mortality 1990-1995. Socioeconomic status was assigned from occupation in 1990 or 1980. Alcohol-related deaths were defined from underlying or contributory causes. Poison regressions were applied to compute age-adjusted mortality rate ratios for all-causes, alcohol-related and other causes among 30-79-year-olds. The contribution of alcohol to mortality differentials was calculated from absolute differences. RESULTS: Around 5% (9,547) of all deaths were alcohol-related (30-79 years). For both sexes, manual workers, lower nonmanuals, entrepreneurs and unclassifiable groups had significantly higher alcohol-related mortality than did upper nonmanuals. Male farmers had significantly lower such mortality. The contribution of alcohol to excess mortality over that of upper nonmanuals was greatest among middle-aged (40-59 years) men who were manual workers or who belonged to a group of 'unclassifiable & others' (25-35%). It was of considerable size also for middle-aged lower nonmanuals (both sexes), male entrepreneurs, female manual workers and 'unclassifiable & others'. Among men, the total contribution of alcohol (30-79 years) was estimated at 16% for manual workers, 10% for lower nonmanuals and 7% for entrepreneurs; and among women, 6% (manual workers, lower nonmanuals) and 3% (entrepreneurs). CONCLUSION: Although deaths related to alcohol were probably underreported (e.g. accidents), alcohol clearly contributes to socioeconomic mortality differentials in Sweden. The size of this contribution depends strongly on age (peak among the middle-aged) and gender (greatest among men).

Language: en