Food frequency questionnaire reproducibility for middle-aged and elderly Japanese

The aim of this study was to examine the reproducibility of a self-administered food frequency questionnaire (FFQ) originally developed for the Japan Public Health Center-Based Prospective Study (JPHC study) and modified for use in the Next Generation (JPHC-NEXT) study.

Participants included 98 men and 142 women aged 40-74 years from the five areas included in the JPHCNEXT protocol. In November 2012, participants were recruited and asked to complete the first nutrition survey. The second nutrition survey was completed after 1 year.

We estimated daily energy as well as 53 nutrient and 29 food group intakes using the FFQ. To assess reproducibility, we calculated Spearman correlation coefficients between both FFQs, which showed mostly intermediate-to-high values. Median (range) correlation coefficients and quartile distribution in the same and adjacent categories for energy-adjusted nutrient intakes were 0.55 (0.42-0.84) and 84.7% (76.5%-98.0%) in men and 0.54 (0.35-0.80) and 84.5% (76.1%-94.4%) in women. The respective values for energy-adjusted food group intakes in men and women were also mostly intermediate to high: 0.54 (0.39-0.79) and 83.7% (75.5%-90.8%) in men and 0.57 (0.40-0.83) and 84.5% (77.5%-93.7%) in women.

The FFQ developed for the JPHC-NEXT study has reasonable reproducibility. Because this FFQ has also been validated in a previous study, it can be considered a useful dietary assessment tool to examine associations between dietary consumption and lifestyle-related diseases.