Validation and reproducibility of a food frequency questionnaire to assess energy and fat intake in affluent north Indians.

ABSTRACT

BACKGROUND: India is currently witnessing a sharp rise in noncommunicable disorders such as obesity, diabetes, hypertension and cardiovascular diseases. This rise can be related in part to dietary changes such as increased intake of calories, fat (especially saturated fat) and cholesterol. A simple, accurate and reproducible method to measure these nutrients is essential to study the role of diet in these diseases in epidemiological studies. We aimed to develop and validate a food frequency questionnaire that could be used for this purpose. METHODS: Thirty urban north Indian subjects (age 23-64 years, 16 men) belonging to a high socioeconomic group were studied. The subjects were selected consecutively over a period of 3 weeks from among those participating in an epidemiological survey on cardiovascular risk factors in an affluent population. A 102-item food frequency questionnaire was developed to capture the intake of calories, fat, saturated fat and cholesterol. The results obtained by the food frequency questionnaire were compared with a 5-day diet record. To assess the reproducibility of the food frequency questionnaire, it was re-administered after 3 months to the 23 subjects available. RESULTS: It took the dietician 20 minutes or less to administer the questionnaire. There was good correlation between the nutrient values as calculated by the food frequency questionnaire and 5-day diet record. The correlation for energy intake was 0.80, and varied between 0.55 and 0.69 for unadjusted intake of other nutrients. After adjusting for calories, the correlation varied between 0.45 and 0.68. In general, the food frequency questionnaire overestimated the energy-adjusted nutrient intake by 6%-17%. When intake was classified into quartiles, there was good agreement between the two methods: 43%-100% for calories; 29%-86% for other nutrients for unadjusted intake; 29%-71% for nutrients after energy adjustment. On calculation of intake after re-administration of the food frequency questionnaire, there was a moderate to strong correlation (energy adjusted r=0.49-0.90) between the two evaluations for various nutrients. CONCLUSION: The food frequency questionnaire developed for the assessment of nutrient intake in a north Indian population was easy to administer, showed moderate to good correlation with the 5-day diet record and was reproducible.