Validation of anthropometric-based weight estimation equations among adults: A cross-sectional study in Kira Health Center IV in Wakiso District, Central Uganda

Abstract

Introduction: Many a time, clinicians are faced with emergencies and they need an accurate estimate of patients’ weight but because of the inadequate resources, especially in resource-poor countries, they neither have the means nor the time to get the patients’ accurate weight to assess and manage the patients appropriately. They resort to using inaccurate methods of weight estimation like the visual estimation of weight. This leads to errors in the diagnosis and management of patients. Using other anthropometric measurements to predict weight has proved to be relatively accurate elsewhere, therefore it can be validated and adopted for use in Uganda in circumstances where it’s difficult to measure patients’ weight. Objectives: To validate the accuracy of other anthropometric measurements in predicting weight in the Ugandan population. Methodology: A cross-sectional study was conducted at Kira Health Center IV. A sample of 240 adult patients, 18 years and above, was selected from among inpatients and outpatients. Their demographic information was obtained and they were subjected to several anthropometric measurements. The continuous data were summarized using mean and standard deviation and the categorical data was summarized using frequency and proportions. The accuracy of the different equations suggested in the literature was determined using paired t-test and Bland Altman analysis. A weight estimation equation was developed using multiple linear regression analysis and validated using paired t-test, Bland Altman analysis, and percentage bias within 5%, 10%, and 20% of the actual weight. Results: A sample of 240 participants was selected by consecutive sampling stratified by sex and body mass index. Using Bland-Altman analysis, equation R3 was found to have the best agreement with the actual weight with a mean difference (standard deviation) of 2.55(6.99). Circumference measurements were found to be the best predictors of weight. The weight prediction equations developed and validated were, “1.5735 x MAC + 0.3604 x AC + 1.0385 x CC – 46.7114” for females and “1.1924 x MAC + 0.6009 x AC + 0.5895 x CC – 38.2433” for males. Weight prediction by the equations was more accurate for females than males. Conclusion: Anthropometric measurements other than weight can accurately predict weight. Among all anthropometric measurements, circumference measurements are the best predictors of weight and a combination of which can be used to optimally predict weight. Therefore in the absence of appropriate weighing scales to weigh patients, clinicians can use the developed anthropometric-based weight estimation equations can act as a relatively accurate alternative.