We examined the reproducibility and the validity of the food frequency questionnaire used in the Latin American Study of Oral and Larynx Cancer. Telephone interviews were carried out with a sample of 35 participants who were controls in that study. The original food frequency questionnaire and a 24-hour recall were applied. The reproducibility was evaluated by the comparison of food consumption described in both food frequency questionnaires, and the validity by the comparison of food consumption in the 24-hour recall and the second food frequency questionnaire. In order to compare questionnaires, we used the Wilcoxon test to measure food consumption differences, the Spearman coefficient to analyze food consumption correlations, the intraclass correlation to estimate intrapersonal variability, and the Kappa test to analyze food consumption agreement. Statistically significant differences in consumption of pasta and vegetables were observed in the reproducibility analysis. The validity analysis showed statistically significant differences for butter, tubercles, chicken meat, vegetables, leguminous and fruits; in general, the food frequency questionnaire over-estimated the consumption of these food groups. Agreement of individuals' classification in tertiles of food consumption ranged from 40% to 80% between the two food frequency questionnaires, and from 31% to 74% between the 24-hour recall and the second food frequency questionnaire. These results revealed that the food frequency questionnaire used in the Latin American Study has good reproducibility and reasonable validity for food consumption.
Food frequency questionnaire; Reproducibility; Validity; Case-control study; Diet and cancer
