Nutritional epidemiology [review]

Epidemiological techniques constitute an important method of nutritional investigation. Historical and geographical evidence suggests hypotheses which can be tested by more detailed studies. Cross-sectional surveys describe populations and provide information about prevalence; they also enable the interrelationships of various nutritional and physiological indices to be investigated. The case-control study is useful in exploring possible aetiological factors in diseases, but has the drawback that past dietary information is difficult to obtain accurately. Prospective studies identify predictors of disease and death; they are especially valuable in elucidating the natural history of disease, and supply good evidence on causation, but their duration and expense are disincentives to their use. Randomized controlled trials provide the most precise evidence about causation, and should be used whenever feasible to answer specific questions