(721) Disparities in the Accuracy of the Broselow Tape in U.S. Children

Introduction: Accurate weight estimation is an essential aspect of resuscitative care in critically ill children. Tools such as the Broselow-Luten System (“Broselow Tape”) have guided early resuscitation efforts when direct weight measurements are not feasible. The accuracy of this tool has not been recently evaluated at a large scale and thus we aimed to determine the accuracy of Broselow weight estimations in the general pediatric U.S. population.

Methods: A retrospective analysis of data from the 1999-2020 National Health and Nutrition Examination Surveys (NHANES) conducted by the CDC was performed. Subject-level data included age, sex, race (Mexican American [MA], Non-Hispanic White [nHW], Non-Hispanic Black [nHB], Other Hispanic [OH], and Other/Multiracial [OM]), weight, height, and body mass index. The inclusion criteria was age ≤ 144 months. Predicted weight based on the Broselow color zone was compared against actual weight and reported as the percent difference with an accurate estimation defined as a difference of ≤ ±10%.

Results: 27,413 subjects (50mo [IQR 20-80], 51% male, 17.4kg [IQR 11.7-25.8], 104.5cm [IQR 83.8-124.7]) were analyzed. BMI (16.4 kg/m2 [15.3-18]) was computed for subjects ≥2yo. 5,283 (19%) subjects had inaccurate weight predictions and there were significant differences when stratified by race; nHW had the least erroneous predictions while OH had the most erroneous predictions (p < 0.05). There were also notable differences in the accuracy of weight predictions when stratified by the different Broselow colored zones with the largest inaccuracies occurring in the two tails (“gray” and “green” zones). In multivariable logistic regression, female sex, race other than nHW, and higher weights were all associated with erroneous weight predictions (p < 0.05).

Conclusions: Our analysis suggests an estimated 1 in 5 children may have >10% discrepancy between their actual and predicted body weight when using the Broselow Tape with varying levels of disparities in the accuracy of the different colored zones. We also found differences in prediction accuracies based on race which may contribute to implicit systemic biases. Revisions to this tool or leveraging technologies to develop newer tools for more accurate anthropometric data predictions are critically needed.