Household and community variations and nested risk factors for diarrhoea prevalence in Southern Malawi : a binary logistic multilevel analysis

This paper examines household and community-level influences on diarrhoeal prevalence in southern Malawi. A Bayesian multi-level modelling technique is used in the estimation of hierarchically built data from a survey of individuals nested within households nested within communities. Households have strong unobserved influence on diarrhoeal illness (s2u ¼ 4.476; 95%CI: 2.081, 6.871). A joint Wald test of significance shows that an individual's age [w24 ¼ 55:921; p ¼ 0:000] and school [w22 ¼ 18:203; p ¼ 0:000] have strong influence on an individual's diarrhoeal prevalence. An individual's history of malarial-like illness also has a strong positive relationship with diarrhoeal prevalence [b ¼ 0.606, p ¼ 0.000]. Household factors that influence diarrhoea include employment status of head of household [b ¼ 70.619, p 5 0.021], maternal age [b ¼ 70.013, p 5 0.003], and size of household [b ¼ 70.669, p ¼ 0.000]. The positive relationship between diarrhoea and malaria-like episodes highlights common risk factors hence the need for common approaches to combat the diseases. Significant household effects underline the importance of household considerations in policy issues