Assuming the "Best-Fit" Shape of the Usual Nutrient Intake Distribution (with nutriR) Does Not Affect Prevalence Estimates of Nutrient Intake Inadequacy: Results from a Dietary Intake Survey in Ivorian School-Age Children.

Background: Common methods to assess nutrient intake inadequacy are the estimated average requirement (EAR) cut-point method and the probability of adequacy (PA). It has been postulated that the shape of the intake distribution can have a significant impact on estimates of inadequacy. The "nutritional intake functions for R" ("nutriR") package has been developed to address this.

Objectives: This study aimed to assess the impact of assuming the "best-fit" shape of nutrient intake distributions (using nutriR) on prevalence estimates of inadequate nutrient intake, as compared with observed nutrient intake distributions with the EAR cut-point or PA method.

Methods: A cross-sectional dietary intake survey was conducted among 443 Ivorian school-aged children (6-12 y) by 24-h recall, with nonconsecutive repeats for 13% of the population. Six methods to estimate the prevalence of intake inadequacy for 11 nutrients were compared: methods 1-3 used single 24-h recall data to either apply the 1) EAR cut-point method, 2) the PA method, or 3) the nutriR method; and methods 4-6 applied the same methods after correction for within-person variation using the Statistical Program to Assess Usual Dietary Exposure, which generates estimates of usual daily nutrient intake.

Results: Nutrient intake distributions were right skewed (skewness average 1.0, range: 0.4-2.0) and showed large variation (coefficient of variation average 40%, range: 20%-60%). The prevalence estimates of inadequate nutrient intake were similar across all methods, with a mean difference of ±5 percent point (range: 0-27 percent point). The smallest mean differences, 0.4 and 0.3 percent points, were observed with usual intake data between nutriR and EAR cut-point (method 6 compared with 4), and nutriR and PA (method 6 compared with 5), respectively. Larger differences (∼12 percent point) were observed when comparing single 24-h recall data (methods 1-3) with usual intake (methods 4-6).

Conclusions: Assuming nutrient intake distributions with the shape of "best-fit" to the observed intake distribution did not affect prevalence estimates of nutrient inadequacy in this population of Ivorian school children. This study was registered at clinicaltrials.gov as NCT06175130.