Impact of Dietary Assessment Methods on Reports of DS Product Use and Nutrient Intakes Among DS Users

OBJECTIVES: Reliable dietary assessment tools for assessing nutrient intake from foods/beverages and dietary supplements (DS) are critical for population-level monitoring; yet, different tools vary in their strengths and weaknesses for assessing intake from foods/beverages, and less is known about the performance of these tools in measuring intake from DS. The Interactive Diet and Activity Tracking in AARP (IDATA) cohort study was designed to evaluate the measurement error structure of 24-hr dietary recall (24 HR) and food-frequency questionnaire (FFQ) data, inclusive of DS information. The objective of the study was to compare DS data reported on 24 HRs and frequency-based questionnaires from the IDATA study among U.S. adults (aged 50–74 y; n = 795). METHODS: DS information was collected over 12 months among adults who participated in IDATA and completed up to 6 Automated Self-Administered 24 HRs (ASA24s; 2011 version) and 2 FFQs (i.e., Diet History Questionnaire (DHQ) II), which queried information on DS usage, frequency, and amount of DS taken. Prevalence of use by product type and estimated nutrient intakes from DS for selected nutrients (i.e., calcium, magnesium, vitamins C and D) were compared between the 2(nd) DHQ II and the average of ≥2 ASA24s using the kappa coefficient to quantify agreement and paired t-tests to compare mean nutrient intakes. RESULTS: The majority of U.S. adult DS users who participated in IDATA reported DS use on the DHQ (84%) and on ≥2 ASA24s (84%). Among the DS products reported on ≥2 ASA24s and the DHQ, agreement between the two methods ranged from κ = 0.06 (multivitamin) to κ = 0.70 (calcium and vitamin D), with substantial agreement present (κ ≥ 0.61) for iron, omega-3, fish oil, multivitamin-mineral, potassium, vitamin C, and vitamin E products. When evaluating differences in mean nutrient intake reported on ≥2 ASA24s and on the DHQ, no significant differences in calcium or vitamin C intake were observed; however, for vitamin D, large variation in intake was observed between the DHQ (8.3 μg/d) and ≥2 ASA24s (19.9 μg/d). CONCLUSIONS: Currently, no consensus exists on the best method for measuring intake from DS. The comparability of 24 HRs and frequency-based dietary assessment methods in measuring the prevalence of use of and estimating nutrient intakes from DS varies by the nutrient of interest, and the type of DS product consumed. FUNDING SOURCES: N/A.