Validation of a food frequency questionnaire for use in Italian adults living in Sicily.

The objective of this study was to validate two interviewer-led food frequency questionnaires (FFQs) of very different lengths: a medium-length FFQ (medium-FFQ) of 36 items and a short-length FFQ (short-FFQ) of 18 items, intending to measure levels of intakes in a local population. Both FFQs were validated against intakes derived from a 3-day dietary record (3-day DR). Sixty-five non-diabetic adults with no known cardiovascular, renal or other systemic diseases were included. High correlation coefficients between the FFQ and the 3-day DR (0.45-0.73) were observed for energy intake, carbohydrates and lipid and protein intake. Bland-Altman plots showed good agreement between the methods. Low (0.26-0.37) correlation coefficients of the different nutrient intakes obtained with the short-FFQ and the 3-day DR were observed, with the exception of alcohol intake (rho = 0.49). This study showed promising evidence for the use of a medium-FFQ as a potentially useful tool for investigating the relationship between habitual diet and diseases in clinical and research settings.