[Review on dietary methods for epidemiologic studies]

Different dietary methods have been used for epidemiologic studies on nutrition, diet and disease. Methods so far reported are classified into dietary record method (including weighing method), recall method, Burke's method, frequency method, and semiquantitative frequency method. The recall method and frequency method are considered to be unsuitable when the average intake of nutrients and foods of individuals is to be assessed. It is difficult for a person to recall food intake over several days, and the 24-hour recall method is unable to evaluate the average intake. Quantitative assessment of intake is hardly possible in the frequency method. Dietary evaluation in case-control studies has limitations in terms of difficulties in retrospective dietary assessment and the change in dietary habit over a long period of time. Current epidemiologic knowledge on diet and disease stimulates cohort studies using the 7-day dietary record or semiquantitive frequency method.