Comparison of a picture-sort food-frequency questionnaire with 24-hour dietary recalls in an elderly Utah population.

Objective: To evaluate the 137-item Utah Picture-sort Food-frequency Questionnaire (FFQ) in the measurement of usual dietary intake in older adults. Design: The picture-sort FFQ was administered at baseline and again one year later. Three seasonal 24-hour dietary recall interviews were collected during the year between the two FFQs. Mean nutrient intakes were compared between methods and between administrations of the FFQ. Setting: The FFQ interviews were administered in respondents' homes or carecentres. The 24-hour diet recalls were conducted by telephone interview on random days of the week. Subjects: Two-hundred-and-eight men and women aged 55±84 years were recruited by random sample of controls from a case±control study of nutrition and bone health in Utah. Results: After adjustment for total energy intake, median Spearman rank correlation coefficients between the two picture-sort FFQs were 0.69 for men aged #69 years, 0.66 for men aged .69 years; and 0.68 for women aged #69 years, 0.67 for women aged .69 years. Median correlation coefficients between methods were 0.50 for men #69 years old, 0.52 for men .69 years old; 0.55 for women #69 years old, 0.46 for women .69 years old. Conclusions: We report intake correlations between methods and administrations comparable to those reported in the literature for traditional paper-and-pencil FFQs and one other picture-sort method of FFQ. This dietary assessment method may improve ease and accuracy of response in this and other populations with low literacy levels, poor memory skill, impaired hearing, or poor vision.