[How valid are self-reported intakes of beer, wine and spirits in population studies?].

In order to compare data on intake of wine, beer and spirits from a frequency questionnaire with intake of each type of alcoholic beverage estimated from a dietary interview, a randomly selected sub-sample of 244 women and 249 men aged 35-65 years was cross-sectionally studied. The sample was a sub-sample of the Danish MONICA study. Mean outcome measure in the study was the differences in intake of beer, wine and spirits as reported by the frequency questionnaire and the diet history interview. We found an overall agreement between the two methods, with very little or no systematic variation for all three alcoholic beverages. We conclude that compared to a more time and money consuming thorough dietary interview, the traditional frequency questionnaires seem to sufficiently capture intakes of different types of alcohol. Bias in alcohol reporting by the frequency questionnaire does not seem responsible for the recently found decreased mortality among subjects with a daily intake of wine, nor the increased mortality from drinking of spirits.