Validation of the Lusaka Formula: A Novel Formula for Weight Estimation in Children Presenting for Surgery in Zambia.

Background: In children, the use of actual weight or predicted weight from various estimation methods is essential to reduce harm associated with dosing errors. This study aimed to validate the new locally derived Lusaka formula on an independent cohort of children undergoing surgery at the University Teaching Hospital in Lusaka, Zambia, to compare the Lusaka formula's performance to commonly used weight prediction tools and to assess the nutritional status of this population.

Methods: The Lusaka formula (weight = [age in months/2] + 3.5 if under 1 year; weight = 2×[age in years] + 7 if older than 1 year) was derived from a previously published data set. We aimed to validate this formula in a new data set. Weights, heights, and ages of 330 children up to 14 years were measured before surgery. Accuracy was examined by comparing the (1) mean percentage error and (2) the percentage of actual weights that fell between 10% and 20% of the estimated weight for the Lusaka formula, and for other existing tools. World Health Organization (WHO) growth charts, mid upper arm circumference (MUAC), and body mass index (BMI) were used to assess nutritional status.

Results: The Lusaka formula had similar precision to the Broselow tape: 160 (48.5%) vs 158 (51.6%) children were within 10% of the estimated weight, 241 (73.0%) vs 245 (79.5%) children were within 20% of the estimated weight. The Lusaka formula slightly underestimated weight (mean bias, -0.5 kg) in contrast to all other predictive tools, which overestimated on average. Twenty-two percent of children had moderate or severe chronic malnutrition (stunting) and 4.7% of children had moderate or severe acute malnutrition (wasting).

Conclusions: The Lusaka formula is comparable to, or better than, other age-based weight prediction tools in children presenting for surgery at the University Teaching Hospital in Lusaka, Zambia, and has the advantage that it covers a wider age range than tools with comparable accuracy. In this population, commonly used aged-based prediction tools significantly overestimate weights.