Reproducibility and relative validity of the food frequency questionnaire for assessing dietary intake among elderly in Northern China.

Abstract

This study aimed to evaluate the reproducibility and relative validity of a food frequency questionnaire (FFQ) for assessing energy and nutrients intake in older adults in Northern China. Dietary data were collected in Tianjin, China from community-dwelling elderly (≥ 65 years) using a 133-item interviewer-administered FFQ. A total of 111 participants completed two FFQs within a 3-month interval to measure reproducibility, and 148 participants completed a 3-day diet record (3DDR) to determine relative validity. The agreement between dietary assessment tools was assessed using correlation coefficients, cross-classification analysis, weighted kappa statistic, and Bland-Altman plots. In the reproducibility study, both energy-adjusted correlation coefficients and intraclass correlation coefficients (ICCs) were above 0.40 (P < 0.05), and weighted kappa scores were above 0.20 for all nutrients. Furthermore, there were no substantial differences in correlation coefficients between subgroups divided by sex. In the validity study, the energy-adjusted correlation coefficients were above 0.2 (P < 0.05) for all nutrients, and weighted kappa scores were above 0.2 for all nutrients except carbohydrate, vitamin E, biotin, zinc, selenium, and manganese. Further subgroup analysis found that the energy-adjusted correlation coefficients for 20 nutrients were higher in males than in females. This developed FFQ is a tool that offers credibly reproducible and relatively valid estimates of energy and certain nutrients for the elderly in Northern China. It could be used to explore the association between dietary intake and health outcomes in future research.