Validation of a food frequency questionnaire to assess macro- and micro-nutrient intake among South Asians in the United Kingdom

Background: The South Asian population is one of the largest minority ethnic groups in the United Kingdom (UK), forming 2.7% of the UK population. Risk of diseases such as CHD, NIDDM is high in South Asians and risk of cancer low in this population compared both to the native UK population and other migrant groups. It is useful to investigate the experience of disease and dietary exposures for aetiological clues in South Asians. The FFQ was designed for a population-based case-control study of diet and breast cancer. Aims: To validate a food frequency questionnaire (FFQ) to assess macro- and micronutrient intake among South Asians in the United Kingdom (UK). Methods: A one-year long study of current diet was conducted using monthly telephone 24-hour recalls followed by administration of an interviewer-administered FFQ to ascertain usual diet during this period. General practices in the Thames and West Midlands regions, England were used to sequentially recruit 100 women from a larger random sample of South Asian migrants from general practitioners' patient lists participating as controls in a case-control study of diet and breast cancer. Results: A total of 133 women were invited to achieve the final sample of 100 (76% response rate). The proportion of individuals classified by the two dietary assessment methods into the same or adjacent quartiles was high ranging from 65% (vitamin A) to 96% (protein). Misclassification into opposite quartiles was very low (0% to 5%), except for vitamin A (10%). Energy-adjusted Spearman correlation coefficients were reasonable for almost all nutrients being highest for protein (0.76), NSP (0.71), folate (0.70) and cholesterol (0.69). Correction for within-person variation in monthly 24-hour recalls had little effect on the magnitude of the nutrient correlations between the FFQ and the 24-hour recalls. Calibration coefficients to correct relative risks for nutrient-disease associations were above 0.50 for most nutrients indicating that the degree of attenuation introduced by the FFQ would be acceptable. Conclusions: This FFQ was specifically designed for South Asian women in the UK. Despite the diversity of diets, the FFQ had reasonable validity. The role of diet in breast cancer disease aetiology in this population is being assessed with this instrument. Sponsorship: The study was supported by Cancer Research UK.