Background: The reproducibility and validity of alcohol consumption has not been adequately studied, particularly in mediterranean countries, where alcohol drinking is a widespread habit, especially during meals.
Methods: We compared alcohol consumption measured by two interviewer-administered food frequency questionnaires (FFQ) with average intake derived from two 7-day dietary (7-DD) records (the reference method) on 395 volunteers. Different types of alcoholic beverages were considered separately in order to verify the ability of the questionnaire to assess detailed patterns of alcohol intake.
Results: A satisfactory level of reproducibility and validity of the pattern of alcohol consumption across different levels and types of alcoholic beverage intake was observed. The reproducibility of wine and total alcohol intake showed correlation coefficients > 0.75 in both sexes. The validity was somewhat higher for wine (around 0.70) than for other alcoholic beverages and total alcohol intake. This is probably accounted for by the more regular pattern of wine consumption during the year as compared to other alcoholic beverages (beer, grappa, etc) which are more strongly influenced by seasonal and daily variations. However, about 30% of abstainers according to FFQ were drinkers by the reference method. The opposite was observed in only 4% of subjects.
Conclusion: The FFQ is a reliable and valid instrument for collecting alcohol intake in regular drinkers. Lower validity in irregular drinkers may be due to seasonal variation and/or inadequacy of the FFQ to capture irregular patterns of consumption and/or inadequacy of the average of two 7-DD as a reference method. Furthermore, a considerable degree of misclassification was observed between non-drinkers and moderate drinkers.