Estimating Exposure in Quantitative Microbial Risk Assessment Models Using Dietary Recall Data

Introduction: Quantitative microbial risk assessment (QMRA) models require information on how often a food is consumed and the amount consumed to characterize the risk from exposure in food. In the United States, the National Health and Nutrition Examination Survey (NHANES) provides comprehensive consumption data, but its use in QMRA is complicated. Data is obtained via two-day dietary recall over a two-year cycle from a subset of the population that oversamples specific sub-populations. Sample sizes from a single cycle are generally adequate, except for sub-groups and infrequently consumed foods. Therefore, multiple NHANES cycles are often used to improve precision in dietary exposure models but this assumes consistent consumption patterns over time. 

Purpose: This study assessed the precision of the proportion of the population reporting consumption of a commodity and the average number of grams consumed per day, obtained with one versus two data cycles.

Methods: Dietary recall data from the 2009-2010 and 2011-2012 NHANES cycles were used to calculate, for individual and combined cycles, point estimates and 95% confidence intervals (CI) for 1) proportion of participants reporting consumption of a food commodity on either day, and 2) average grams consumed per day for days the commodity was consumed. Precisions of the estimates were compared to evaluate potential bias and variability.

Results: Minimal gains in precision were seen using two versus one cycle of NHANES data. For foods consumed by 0.1% of the population, the 95% CI were (0.4-0.19) and (0.06-0.16) for one and two cycles, respectively. Similarly, the estimated grams consumed per day for baked potatoes, consumed by 5% of the population, was 144.2 g/day (95% CI: 137.4-151.0) for one cycle and 147.0 g/day (95% CI: 141.5-152.5) for two cycles.  

Significance: Improving the precision of dietary exposure estimates for QMRA is critical to providing actionable information to decision-makers.