Evaluation of a diet history questionnaire for epidemiologic studies.

This study was conducted to evaluate a detailed diet history questionnaire. Diet histories obtained from 16 men were compared with food records kept by their partners for a 30-day period. Mean estimates for all 13 nutrients calculated were higher with the diet history than with food records. Positive correlation coefficients were found between estimates from the two methods for all nutrients as well as food groups and six of the 13 nutrients gave a good fit for the corresponding regression equations. The questionnaire was also evaluated by repeating it after six months on 26 case-control pairs of an ongoing diet and cancer study. High correlation coefficients were obtained for most nutrients for the controls. However, cases showed lower correlations probably due to changes in their diet. The study found that the diet history method as an estimate of dietary patterns among groups shows sufficient validity and reliability to make it a useful instrument for epidemiologic studies.