Abstract
BACKGROUND: Food frequency questionnaires (FFQs) are dietary surveys widely used in studies of nutritional epidemiology because they are inexpensive, easy to complete and provide significant information on dietary intake over a long period of time. FFQs should be written specifically for each study group since diet may be influenced by ethnicity, culture, economic status and environmental factors. The aim of the authors on this work was to design and validate a food frequency questionnaire for the Peruvian Amazon population. METHODS: Three hundred three individuals were selected and completed three 24-h recalls (R24). Two FFQs were conducted on two different occasions (FFQ.1 and FFQ.2). The validity of the FFQ was evaluated by comparing the relationship between the average daily nutrient intake estimated by the FFQs and that obtained by the three R24. The reproducibility was measured by comparing the mean nutrient intake of the two FFQs carried out. The correlations were assessed using the Pearson correlation coefficient, the intraclass correlation coefficient (ICC), the Lin correlation coefficient (CCC) and the Bland-Altman plot. RESULTS: The results obtained to establish the validity showed a high correlation, with an average Pearson's correlation coefficient of 0.70, a CCI of 0.65 and a CCC of 0.60. Approximately 60% of the nutrients had a CCC above 0.60. In terms of reproducibility, better results were obtained, with an average Pearson's correlation coefficient of 0.67, 0.64 for CCI and 0.58 for CCC. CONCLUSIONS: The correlation coefficients show good validity and reproducibility, and therefore, the FFQ we have developed may be considered a useful and valid tool to estimate the dietary intake of the Peruvian Amazon population.