Evaluation of the Relative Validity of the Short Diet Questionnaire for Assessing Usual Consumption Frequencies of Selected Nutrients and Foods.

A 36-item Short Diet Questionnaire (SDQ) was developed to assess usual consumption frequencies of foods providing fats, fibre, calcium, vitamin D, in addition to fruits and vegetables. It was pretested among 30 community-dwelling participants from the Québec Longitudinal Study on Nutrition and Successful Aging, "NuAge" (n = 1793, 52.4% women), recruited in three age groups (70 ± 2 years; 75 ± 2 years; 80 ± 2 years). Following revision, the SDQ was administered to 527 NuAge participants (55% female), distributed among the three age groups, both sexes and languages (French, English) prior to the second of three non-consecutive 24 h diet recalls (24HR) and validated relative to the mean of three 24HR. Full data were available for 396 participants. Most SDQ nutrients and fruit and vegetable servings were lower than 24HR estimates (p < 0.05) except calcium, vitamin D, and saturated and trans fats. Spearman correlations between the SDQ and 24HR were modest and significant (p < 0.01), ranging from 0.19 (cholesterol) to 0.45 (fruits and vegetables). Cross-classification into quartiles showed 33% of items were jointly classified into identical quartiles of the distribution, 73% into identical and contiguous quartiles, and only 7% were frankly misclassified. The SDQ is a reasonably accurate, rapid approach for ranking usual frequencies of selected nutrients and foods. Further testing is needed in a broader age range.