Weight estimation in paediatric resuscitation: A hefty issue in New Zealand

To test the accuracy of weight estimation methods currently used in New Zealand to predict a child's weight in emergency resuscitation.

A prospective, observational study. Data were collected in July 2013 at five Auckland schools among children aged 5-10 years. Collected demographic information included age, ethnicity, sex and school decile. Standardised measures included weight, height and Broselow-Luten tape (2011 version, limited to children 43-143 cm) weight. Age-based weight estimates were calculated for APLS, Shann and Theron formulae. Mean bias (actual weight - estimated weight) and clinical accuracy (proportion of estimates within 10% of actual weight) are reported. Bland-Altman plots illustrate agreement and 95% limits of agreement.

The 376 participants weighed between 14.2 and 93.1 kg. The proportion of weight estimates within 10% of actual weight were 28.7%, 39.1% and 45.7% for the age-based formula (Theron, APLS and Shann, respectively). The mean bias was negative for Theron (-6.5) and positive for APLS (7.8) and Shann (7.7). For the length-based Broselow-Luten tape method (n = 305), the proportion of weight estimates within 10% of actual weight was 73.4% and mean bias was 1.1.

For children under 143 cm in height, the Broselow-Luten tape outperforms other weight estimation methods, accurately estimating weight in approximately three out of four children. The age-based estimation methods performed poorly overall, with variation by age and ethnicity.