A model for habitus-adjusted paediatric weight estimation by age and data concerning the validation of this method on a large dataset of English children

It is often not possible to weigh children upon arrival at an emergency room before commencing the provision of emergency care. Because drugs for children are prescribed on the basis of age and body weight, estimations of weight are necessitated. Age-based equations have been one of the most commonly used weight estimation strategies historically. Due to the variability of weight for age in children, and variations in body habitus, these methods are inaccurate by design (Young and Korotzer, 2016) [1].


Subject area
Resuscitation Science More specific subject area

Paediatric weight estimation
Type of data

Value of the Data
Provide a model for assessment of this method in other populations Aid researchers in understanding the extent to which age-based weight estimation strategies may be improved by taking into account a habitus assessment Serve as a methodological and performance benchmark for further attempts to optimise paediatric emergency weight estimation by age.

Data
A Bland Altman plot of the performance of the Helix Method on the NCMP dataset. An exported model of the Helix method which may be used by other research groups to interrogate this technique in different populations.

Experimental design, materials, and methods
As described in the original paper (Resuscitation, in press), the NCMP and CDC data were processed using the Python programming language (Python Software Foundation. Python Language Reference, version 3.6), and the pandas library (version 0.20.2, http://pandas.pydata.org), in the Jupyter Notebook computational environment (https://jupyter.org).
The Helix method uses 7 habitus scores, each mapping closely onto the habitus scores used by Wells et al. in the PAWPER-XL system [4]. We assigned CDC weight-for-age and BMI-for-age centiles to each habitus score as follows: HS1 ¼ 10th, HS2 ¼25th, HS3 ¼50th, HS4 ¼ 75th, HS5 ¼90th, HS6 ¼95th, and HS ¼97th centiles. Each child in the dataset was allocated to the CDC BMI-for-age centile (10th, 25th, 50th, 75th, 90th, 95th, 97th) closest to their actual BMI from the NCMP dataset, and the corresponding CDC weight-for-age centile was used to determine an estimated weight.
The overall accuracy of the model was tested by calculating the mean percentage error as a measure of estimation bias. 95% limits of agreement were calculated as a measure of precision. Statistical analysis was performed using Stata (Stata-Corp. 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP). Detailed understanding of the performance of this model may be gained by reference to the original publication.
The table attached in the repository folder shows the estimated weights for an individual of a given age in months and gender, according to each of the seven habitus scores developed in our model, and which was used in the validation paper (Resuscitation, in press). The BMI values that pertain to the lower and upper limits of the habitus score are included for reference as Table 1. The estimates from the new APLS and EPLS formulae for given ages are also included for reference, accompanying the model description.
The two graphs ( Fig. 1) demonstrate the spread of the estimated weights for each of the seven habitus scores between the ages of 24 months and 144 months in dark grey. Superimposed are the weight estimates according to the APLS and EPALS formulae. The graphs demonstrate that in both boys and girls, the current widely used formula based strategies significantly underestimate the weight of older children and fail to account for the broadening of the weight distribution with age.  These charts make no assumptions about accuracy, but show how they compare numerically. Fig. 2 shows a Bland Altman plot for the Helix method.