Reproducibility and validity of a food frequency questionnaire among residents in the Pearl River Delta, South China.

A localized food frequency questionnaire (FFQ) was developed to capture habitual dietary intake in the Pearl River Delta population. This study aimed to evaluate the reproducibility and validity of this FFQ. A total of 219 adults were enrolled in the Pearl River Delta region. In the reproducibility analysis, the participants completed the FFQ twice at a 18-month interval. Validity was assessed using the 24-hour dietary record collected over 9 days across 3 seasons (spring, summer, and winter). For the reproducibility, the ranges of correlation coefficients and kappa values were 0.38 to 0.80 (P < .001) and 0.29 to 0.50 (P < .050), respectively. The consistency for the same and adjacent quartiles was over 80% for most nutrients (87%). For the validity, the ranges of correlation coefficients of the crude, energy adjusted, and kappa values between FFQ and 24-hour dietary record were 0.16 to 0.61, 0.16 to 0.39, and 0.25 to 0.36, respectively, and all these values were statistically significant (all P < .050). The consistency for the same and adjacent quartiles was over 70% for most nutrients (91%). The FFQ used in this study demonstrated acceptable reproducibility and validity (except for fiber, vitamin A, vitamin C, and folate), making it a practical instrument for ranking long-term dietary nutrient intake among adults in the Pearl River Delta, South China.