Background: One-half of US adults report using a dietary supplement. NHANES has traditionally assessed dietary supplement use via a 30-d questionnaire but in 2007 added a supplement module to the 24-h dietary recall (24HR).
Objective: We compared these 2 dietary assessment methods, examined potential biases in the methods, and determined the effect that instrument choice had on estimates of prevalence of multivitamin/multimineral dietary supplement (MVMM) use.
Methods: We described prevalence of dietary supplement use by age, sex, and assessment instrument in 12,285 adults in the United States (>19 y of age) from NHANES 2007-2010.
Results: When using data from the questionnaire alone, 29.3% ± 1.0% of men and 35.5% ± 1.0% of women were users of MVMMs, whereas data from the 24HR only produced prevalence estimates of 26.3% ± 1.1% for men and 33.2% ± 1.0% for women. When using data from both instruments combined, 32.3% ± 1.2% of men and 39.5% ± 1.1% of women were classified as MVMM users. Prevalence estimates were significantly higher by 2-9% in all age-sex groups when using information from both instruments combined than when using data from either instrument individually. A digit preference bias and flattened slope phenomenon were observed in responses to the dietary supplement questionnaire. A majority (67%) of MVMMs were captured on both instruments, whereas 19% additional MVMMs were captured on the questionnaire and 14% additional on the 24HR. Of those captured only on the 24HR, 26% had missing label information, whereas only 12% and 9% of those captured on the questionnaire or both, respectively, had missing information.
Conclusions: Use of both the dietary supplement questionnaire and the 24HR can provide advantages to researchers over the use of a single instrument and potentially capture a larger fraction of dietary supplement users.
Keywords: NHANES; dietary assessment; dietary supplement; measurement error; multivitamin multimineral.
© 2015 American Society for Nutrition.