Association Between Self-Reported Dietary Intake Questionnaires and Objective Measures in an Inpatient Cross-Sectional Study.

Background/Objectives: Measuring dietary intake through self-reported questionnaires can be inaccurate and influenced by sex, eating behavior, and the environment. Here, we compare self-report dietary intake questionnaire responses to objectively measured ad libitum dietary intake in a large, diverse population, and assess differences by sex and food-group composition. Methods: In our inpatient study, from 1999 to 2023, (n = 279) participants completed three different questionnaires assessing different aspects of food intake. Each questionnaire contained the same 77 food items belonging to one of six groups. Groups were either high-fat (HF) or low-fat (LF), then high complex carbohydrate (HCC), high protein (HP), or high simple sugar (HSS). Intake was measured based on the average percent group (PctGrp) intake over three days of ad libitum intake. General linear models, adjusted for relevant covariates and a PctGrp by sex interaction, assessed the relationship between PctGrp intake and questionnaire scores. Results: We found a weak positive correlation between PctGrp intake and food rating (all r ≤ 0.25). There was an interaction between LF/HP and LF/HCC with sex (significant slopes in males only, p = 0.0078, p ≤ 0.0001, respectively). Conclusions: This large study demonstrated little association between self-report dietary questionnaires and intake, especially in females with regards to low-fat foods.