Bayesian modeling of quantiles of body mass index among under-five children in Ethiopia

Abstract

Abstract Background Body Mass Index (BMI) is a measurement of nutritional status, which is a vital pre-condition for good health. The prevalence of childhood malnutrition and the potential long-term health risks associated with obesity in Ethiopia have recently increased globally. The main objective of this study was to investigate the factors associated with the quantiles of under-five children’s BMI in Ethiopia. Methods Data on 5,323 children, aged between 0-59 months from March 21, 2019, to June 28, 2019, were obtained from the Ethiopian Mini Demographic Health Survey (EMDHS, 2019), based on the standards set by the World Health Organization. The study used a Bayesian quantile regression model to investigate the association of factors with the quantiles of under-five children’s body mass index. Markov Chain Monte Carlo (MCMC) with Gibbs sampling was used to estimate the country-specific marginal posterior distribution estimates of model parameters, using the Brq R package. Results Out of a total of 5323 children included in this study, 5.09% were underweight (less than 12.92 BMI), 10.05% were overweight (BMI: 17.06 – 18.27), and 5.02% were obese (greater than or equal to 18.27 BMI) children’s. The result of the Bayesian quantile regression model, including marginal posterior credible intervals (CIs), showed that for the prediction of the 0.05 quantile of BMI, the current age of children [$$\upbeta$$ β = -0.007, 95% CI :(-0.01, -0.004)], the region Afar [$$\upbeta$$ β = - 0.32, 95% CI: (-0.57, -0.08)] and Somalia[$$\upbeta$$ β = -0.72, 95% CI: (-0.96, -0.49)] were negatively associated with body mass index while maternal age [$$\upbeta$$ β = 0.01, 95% CI: (0.005, 0.02)], mothers primary education [$$\upbeta$$ β = 0.19, 95% CI: (0.08, 0.29)], secondary and above [$$\upbeta$$ β = 0.44, 95% CI: (0.29, 0.58)], and family follows protestant [$$\upbeta$$ β = 0.22, 95% CI: (0.07, 0.37)] were positively associated with body mass index. In the prediction of the 0.95 (or 0.85?) quantile of BMI, in the upper quantile, still breastfeeding [$$\upbeta$$ β = -0.25, 95% CI: (-0.41, -0.10)], being female [$$\upbeta$$ β = -0.13, 95% CI: (-0.23, -0.03)] were negatively related while wealth index [$$\upbeta$$ β = 0.436, 95% CI: (0.25, 0.62)] was positively associated with under-five children’s BMI. Conclusions In conclusion, the research findings indicate that the percentage of lower and higher BMI for under-five children in Ethiopia is high. Factors such as the current age of children, sex of children, maternal age, religion of the family, region and wealth index were found to have a significant impact on the BMI of under-five children both at lower and upper quantile levels. Thus, these findings highlight the need for administrators and policymakers to devise and implement strategies aimed at enhancing the normal or healthy weight status among under-five children in Ethiopia.