OBJECTIVE: To assess the relative validity of a Quantitative Food Frequency Questionnaire (QFFQ) for a case-control study on dietary factors and breast cancer in the state of Paraíba, Northeastern Brazil. METHODS: The validity of a 68-item QFFQ was investigated in 38 women, ranging from 25 to 80 years old, potential controls for a case-control study. Four 24-hour dietary recalls (24HDR) were administered by an interviewer, by telephone, during the 1-year period, and a QFFQ was applied at the end of the period to compare mean nutrient intakes. Paired t-tests were applied to verify differences in the mean of nutrients in both methods. Pearson correlation coefficients on nutrient data before and after energy adjustment in both methods were calculated and within-person variability was corrected for (de-attenuated correlation coefficient). RESULTS: Relative validity was verified for energy, carbohydrates, lipids, protein, Vitamin A and Vitamin C. Satisfactory values for energy (r=0.53) and lipids (r=0.40) were encountered for Pearson crude correlation coefficients, but remarkably lower for Vitamin C (r=0.08). When coefficients were corrected for within-person variability, all nutrients presented considerably higher values (r= 0.36 0.87). CONCLUSIONS: Results suggest that QFFQ presents a reasonably valid measure of intake for most nutrients and can be used for diet and cancer studies in women in the State of Paraíba, northeastern Brazil. Nevertheless, calibration procedures will be required for utilization in different geographical areas of Brazil .
Validity; Food frequency questionnaire
