Monitoring dietary change in populations and the need for specific food targets; lessons from the North West Thames Regional Health Survey

Dietary goals are usually expressed in the form of consumption of absolute quantities of specific nutrients. Translating such information into food patterns is difficult. Consequently it is not easy to convey these goals in the promotion of a healthy diet or to monitor dietary change in populations. This paper explores the difficulties and advantages of a food frequency questionnaire approach as used in a large regional health survey to describe food patterns and hence to monitor change in relation to dietary variables. There was considerable variation in the frequency of consumption of individual foods between different sub-groups in the population and these were as expected. To assess the overall diet a composite dietary index was proposed. The index score was significantly associated with various socio-demographic and lifestyle variables. Unhealthy lifestyles were associated with low scores (i.e. an ‘unhealthy’ diet) on the index indicating a measure of validity. However there were several limitations in the food frequency questionnaire method used: neither the questionnaire nor simple global indices of a ‘healthy’ diet derived from the questions used had been validated against a more robust quantifiable measure of diet; some of the questions were not specific for population groups of interest; some questions grouped foods in categories that masked differences in consumption and several did not correspond to recent food targets; and finally there was no energy adjustment, which meant that the obese appeared to have a healthy dietary pattern. The use of a valid set of food frequency questions as an efficient and practical way of monitoring a population's diet by survey is suggested. The questions must be validated against reliable measures of nutrients, and be compatible with food targets based on nutritional recommendations.