Relative validity and reproducibility of a food frequency questionnaire designed for French adults

We tested the reproducibility and relative validity of a French self-administered food frequency questionnaire (FFQ) against 24-hour dietary records (DRs).

We selected 140 participants from the SU.VI.MAX study. They filled in the FFQ twice at 1-year intervals and completed monthly 24-hour DR during that year. The number of food items in the FFQ was 240.

Reproducibility was high. Medians of Pearson coefficients were 0.70 and 0.65 for nutrients in men and women, respectively. Medians of intraclass correlation coefficients for food were 0.62 in men and 0.65 in women. For nutrients, crude Pearson correlation coefficients for relative validity ranged from 0.28 to 0.67 in men and from 0.25 to 0.55 in women. The Spearman coefficient for alcohol was 0.78 in both genders. For food, the medians of the Spearman coefficient were 0.49 and 0.45 for men and women, respectively. Complete disagreement between the FFQ and 24-hour DRs for ranking subjects according to food consumption was low (about 4%), except for rarely consumed food groups, that is nuts, legumes and offal. Calibration coefficients were estimated by linear regression. All coefficients were lower than one.

Acceptable reproducibility and relative validity were observed in a FFQ designed for French adults.