Reproducibility and validity of a food frequency questionnaire in European and Polynesian New Zealanders

The reproducibility and validity of a self-administered 142-item food-frequency questionnaire (FFQ) was assessed in a population comprising 124 European and 52 Polynesian (17 Maori and 35 Pacific Island) New Zealanders aged 40-65 years. Reproducibility correlation coefficients, determined by administration of the same questionnaire on two occasions 3 years apart, were higher in European than Maori and Pacific Island participants, ranging from 0.47 to 0.87 in Europeans (median 0.66) and from 0.41 to 0.79 in Maori and Pacific Island people (median 0.44). In general, there were no significant differences in mean nutrient intakes calculated from the two FFQs by Europeans or Maori and Pacific Island participants despite their cultural and language differences. When the FFQ was compared with a 3-day food diary in a subsample of 101 Europeans, 15 Maori and 22 Pacific Islanders, the validity was good for most nutrients, with overestimation of a few nutrients in each ethnic group. Correlation coefficients between the 3-day food diary and FFQ ranged from 0.41 to 0.81 in Europeans (median 0.48) and from 0.36 to 0.56 in Maori and Pacific Island people (median 0.55). Ratios of energy intake to resting metabolic rate suggested that Maori and Pacific Island people were more likely to underestimate their habitual energy intake by the 3-day diet diary method compared to Europeans, but that Europeans were more likely to underestimate total energy intake by the food frequency method and Pacific Island participants to overestimate it. Obese Europeans and Maori were more likely to under-report dietary intakes by the 3-day diary method. We conclude that our FFQ performed better in European than Maori and Pacific Island participants.