Development and relative validity of a self-administered semi-quantitative drinks frequency questionnaire, validated in a population of university students

Aim Drinks can contribute to energy as well as micro- and macro-nutrient intake. Drink-specific, validated, quick and easy to administer questionnaires are needed to monitor secular trends in beverage consumption for different populations. The aim of the present cross-sectional study was to develop and validate a drink-specific questionnaire for the Greek population. Methods A 41-item semi-quantitative drinks frequency questionnaire (DFQ) was developed and validated against 7-day weighed food records, among young Greek adults. Wilcoxon ranks tests, Kendall tau-b correlations and Bland-Altman plot were used to compare the two methods. Fifty-nine Greek University students (28 males, 31 females) completed both methods. Results No difference was observed in total drinks intake between the two methods. The Bland-Altman plot showed that the two methods are likely to agree concerning the total drinks intake. Wilcoxon's rank test showed no significant differences between the two methods, except for the 'whiskey/vodka/gin' intake (P < 0.001). 'Good' agreement (tau-b > 0.61) was observed for most DFQ items, including low-fat and chocolate milk, hot chocolate, Greek/espresso coffee, commercial tea, packaged fruit juices, wine, alcohol-free beer, light cola-type drinks and water (tap/bottled/sparkling). 'Low' agreement (<0.30) was noted for filtered/instant/iced coffee, freshly squeezed orange juice and some alcoholic drinks (beer/whiskey/vodka/gin). Conclusions The DFQ appears to be an overall valid means of estimating the habitual intake of drinks in large-scale epidemiological surveys. The study supports the use of DFQ for assessing drinks consumption patterns and secular trends.