Collection of food intake data: an evaluation of methods.

Using USDA's recently completed methodological survey designed to evaluate longitudinal measures of food intake, the authors assessed nine methods for measuring dietary status. Methods evaluated included various combinations of prospective and retrospective recording of daily intake; personal, telephone, and mail- and telephone-assisted reporting; semistructured and open-ended recording documents; and number and spacing of days for recording intakes. Results from tabular and regression analyses indicated that the most effective and least costly methods for assessing mean intakes for groups of individuals were those using telephone and/or personal contact and collecting 24-hour recalls for several selected nonconsecutive days during the survey period. More specifically, it was concluded that the method including an initial personal interview for 24-hour recall followed by three telephone interviews for 24-hour recalls over a 1-year period and the method including four telephone interviews for 24-hour recalls for a 1-year period provided food energy intake data comparable with the data collected with other methods and with less effort or cost. Additional analyses were recommended to determine the numbers of days required to obtain estimates of usual intake for individuals and for diet components other than food energy.