News - Journal Publication by Tsirizani Kaombe

The Biostatistics programme is pleased to announce latest publication from masters thesis by its 2010 cohort graduate, Tsirizani Kaombe titled "Modelling Distribution of Under-Five Child Diarrhoea across Malawi". The paper has been published on 25 March 2016 with the Journal of Mathematics and System Science by David Publishing Company through both print and online modes, and the online version can be accessed through the link http://www.davidpublisher.org/home/Journal/JMSS with DOI:10.17265/2159/5291/2016.03.001.

The paper compared consistency of estimates obtained using Bayesian semiparametric methods to those from two parametric models of Poisson and Logistic when analysing child diarrhoea data and it used the 2006 Malawi Multiple Indicator Cluster Survey Data. Inference in Bayesian method was done through MCMC while in parametric models it was via iterative Maximum Likelihood. The Bayesian model was fit in BayesX software while STATA was used for parametric models. The paper used Credible Intervals and Confidence Intervals to compare Bayesian and parametric model estimates, respectively while Chi-Square's Goodness of Fit test was used to compare the two parametric models. The results showed complete agreement in direction and significance of the estimates from the two groups of models. Both models picked child age, breastfeeding status, region of stay, and toilet sharing status as significant risk factors and both ruled out effect of mother education, area of residence, and source of drinking water on child diarrhoea. The paper concludes that both Bayesian and parametric models are equally helpful when estimating the child diarrhoea and recommends applying Bayesian in parallel to parametric model as a simple checking tool for reliability of estimates in case parametric model assumptions were not fully met on the data.

Tsirizani Kaombe has published the paper alongside his masters supervisor, Dr. Jimmy Namangale and joins the list of former students who have published their masters thesis work in the programme. The list includes Masford Banda, Alfred Ngwira, Owen Mtambo, Michael Chipeta, James Chirombo, and Bruce Phiri, all supervised by Assoc. Prof. Lawrence Kazembe. The programme encourages its students and graduates to consider publishing their thesis work with International Journals once the thesis has been passed by examiners. We join the Biostatistics family in congratulating Tsirizani Kaombe and his masters supervisor Assoc. Prof. Jimmy Namangale for the achievement.