Validity and reproducibility of the NORBAGREEN food frequency questionnaire

Objective: To measure the validity and reproducibility of the NORBAGREEN food frequency questionnaire (FFQ). Subjects/methods: In Finland, 125 subjects aged 25-64 years sampled from the five main regions of the FINDIET 2002 Study and in Lithuania, 99 citizens of Kaunas aged 19-75 years participated in the study. Reference methods for the FFQ were two 3-day FFQs in Finland and four 24-h recalls in Lithuania. The FFQ was repeated after 6-8 months in both countries. The outcome of the FFQ1 was correlated with the outcome of the reference methods and with the outcome of repeated FFQ2. Cross-classification of food intakes by FFQ1 and the reference methods was examined in tertiles. Results: Validity correlations (FFQ vs the reference method, Spearman's correlation) were for vegetables, fruit and bread 0.50 (P<0.01), 0.53 (P<0.01) and 0.54 (P<0.01) in Finland; and 0.55 (P<0.01), 0.31 (P<0.01) and 0.51 (P<0.01) in Lithuania, respectively. Correlations were smaller for potatoes and fish. The overall proportion categorized in the same or adjacent intake tertiles with the two instruments was over 83% in both countries. Reproducibility correlations varied between 0.51 and 0.75 in the Finnish study, and between 0.51 and 0.83 in the Lithuanian study. Conclusions: The NORBAGREEN FFQ can be used to rank subjects according to vegetable, fruit and bread consumption. Questions on fish and potato consumption need to be developed further.