RESUMEN
A Food-Frequency questionnaire (FFQ) is the method chosen for estimating dietary intake in epidemiological studies because it can provide valid and reliable estimates of usual intake in a variety of populations. Colorectal cancer is associated with diet but no information about validity and reproducibility of dietary exposure assessment is available in Argentina. The purpose of the study is to evaluate the validity and reproducibility of the FFQ used in assessing the relationship between dietary intake and colorectal cancer. Selected nutrient intake, measured by two interviewer-administered FFQs, was compared with average intake derived from using four 24 hour dietary recalls (DR) as the reference method, for 62 control subjects from Córdoba, Argentina. Pearson's product-moment and Spearman correlation coefficients were used. Nutrient intakes were adjusted for energy, sex and age using tertiles of residuals from regression models. Weighted and unweighted kappa statistics were calculated for both FFQs versus DR. The mean correlation coefficients between DR and FFQs were 0.74. Agreement measurement (Kappa) varied from 0.51 to 0.74 for FFQ1 vs. FFQ2 to between 0.71 and 0.87 for FFQ1 and FFQ2. This study demonstrates an acceptable validity and reproducibility for this FFQ. The pattern of the correlation coefficients found in the present study indicates that the FFQ can provide an acceptable assessment of long-term dietary intake in the Argentine population and in other Latin-american areas.