Aims: The preventive paradox prevails if the majority of alcohol problems accrue to the lesser-drinking majority of population, not to heavy drinkers. Evidence for the paradox has been criticized for being based on self-report. The aim was to examine whether the paradox also applies to deaths and hospital admissions.
Design: Data from four surveys representing the Finnish population aged 15-69 years in 1969, 1976, 1984 and 1992 were pooled; those from 1969, 1976 and 1984 (n = 6726) to study alcohol-related hospital admissions and alcohol-related deaths, and those from 1984 and 1992 (n = 5558) to study self-reported problems. The former data were linked with register data on hospital admission and death up to the end of 2002.
Methods: Comparisons were made separately for men and women (1) between the 10% of population with the highest average alcohol consumption and the remaining 90% of drinkers and (2) between those who reported and those who did not report drinking to intoxication.
Results: A total of 3025 men and 2693 women were available for the study of self-reported problems and 2945 men and 2615 women for deaths and hospital admissions. Seventy per cent of all self-reported problems, 70% of alcohol-related hospitalizations, 64% of alcohol-related deaths and 64% of the premature life-years lost before the age of 65 occurred among the 90% of men consuming less. The respective figures for women were 64%, 60%, 93% and 98%. Drinking five or more drinks per occasion was related to more harm than not drinking that much.
Conclusions: In men, the "prevention paradox" appears to apply to a broadly similar degree to hospitalizations and deaths as self-report alcohol-related problems; in women the phenomenon was apparent to a greater degree for deaths than for other markers of harm.