(P19-042-23) Validation of a Food Frequency Questionnaire With 24-Hour Dietary Recalls Among Middle-Aged Overweight and Obese Individuals

Objectives: The validation of food frequency questionnaires (FFQs) by a reference method is an essential step in the use of the instrument in collecting dietary data. Estimates of nutrient intake may be underestimated in individuals with overweight/obesity. The objective of this study is to validate a semiquantitative 166-item web-based FFQ with 24-hour dietary (24HDR) recalls in a group of overweight/obese adults.

Methods: A 166-item web-based FFQ was administered to all participants in a dietary intervention trial, including 43 adult males (42%) and post-menopausal females (58%) (~40-70 years of age) with overweight/obesity. Four unannounced 24HDRs, (3 weekdays and 1 weekend day) were collected with Nutrition Data System for Research (NDSR) 2021 software using the multiple-pass method at months 1, 2, 3 and 4. The FFQ data was collected two weeks after the last 24HDR was collected and was used to estimate the previous 5 months of intake during the trial. Crude and deattenuated Pearson correlations were calculated between the 24HDRs and FFQ. Quartiles of exact percent agreement were calculated between the FFQ and 24HDRs.

Results: The estimated mean intake of calories, carbohydrate, protein, fat, dietary fiber, MUFA, and vitamin E based on the FFQ were 2198 kcals, 246 g, 82 g, 101 g, 30 g, 41 g, and 30 g, respectively. The estimated mean intake of energy, carbohydrate, protein, fat, dietary fiber, MUFA, and vitamin E based on the 24HDRs were 1845 kcals, 201 g, 76 g, 84 g, 21 g, 34 g, and 15 g, respectively. The deattenuated correlations between the FFQ and 24-hour dietary recalls were 0.33, 0.38, 0.32, 0.41, 0.43, 0.48, and 0.82 for energy, carbohydrate, protein, fat, fiber, MUFA, and vitamin E, respectively. All the correlations were significant except for protein. Exact percent agreement between the FFQ and 24HDRs ranged from 28 to 42% and was highest for MUFA. Gross misclassification was less than 5% for most nutrients, except for carbohydrate and fiber, both of which were 7%.

Conclusions: This web-based semiquantitative FFQ may provide a fairly good estimate of macronutrient and vitamin E intake, with minimal gross misclassification, in overweight/obese individuals. Nutrient estimates were higher based on the FFQ than the 24HDRs.

Funding Sources: None