Validity and reproducibility of a quantitative food frequency questionnaire for a cohort study in Japan

A self-administered quantitative food frequency questionnaire (Qx) was developed for a population-based cohort study on cancer in Takayama, Japan.

The Qx was tested among 58 male and 59 female volunteers. Average daily nutrient intakes for the previous year calculated from the Qx were compared with those from 3-day food records and four 24-h recalls. The Qx was also validated among 37 volunteers by comparing the nutrient intakes calculated from the Qx with 12 1-day food records during a year. We also calculated the intra-class correlation coefficients for various nutrients between the Qx and the second Qx administered by the same volunteers 1 year after the first survey.

Pearson correlation coefficients between total energy from the Qx and 3-day records were 0.38 for men and 0.25 for women and those between the Qx and 24-h recalls were 0.19 and -0.02 for men and women, respectively. Correlations between the several nutrients from the Qx and 3-day records ranged from 0.2 to 0.5 for both men and women. These correlations after energy adjustment ranged from 0.2 to 0.6 for men and from 0.1 to 0.7 for women. In general, the correlations for various nutrients between the Qx and 12 1-day records were higher than those described above. The intra-class correlation coefficients ranged from 0.46 to 0.78 in men and from 0.36 to 0.67, except for vitamin C in women. When the information on portion size was excluded, almost all of the above indices showed somewhat lower figures.

These results suggest that our food frequency questionnaire with portion size information can be used to estimate nutrient intakes of each individual.