Modelling of Malaria Risk Factors in the Mpohor District of Ghana using Logistic Regression

Abstract

This study was aimed at assessing and deriving a predictive model for the relationship between malaria prevalence and malaria causing factors (covariates) in the Mpohor District (which is located in the Western Region of Ghana) by logistic regression. Risk factors such as seasonality (wet or dry), altitude, mining community, proximity of water body, vegetation proximity and clinic proximity were assessed using logistic regression. Collinearity test was performed to avoid information duplicate and multicolinearity by examining the Variance Inflation Factor (VIF) of each covariate. The relationship between malaria and its underlying factors was analysed through stepwise logistic regression where the wald statistics and odds ratio (OR) proved their significance. The results showed that the risk factors such as altitude, seasonality, water body proximity, vegetation proximity and mining community were significant predictors of malaria morbidity in the District . However, it was found that proximity of health facility to community was not a good malaria morbidity predictor. It was recommended among other things, that further research involving more communities in the District and including other known malaria factors be carried out to provide complete and more reliable information that is useful in malaria control. Keywords: Mslaria prevalence, modelling ,odds ratio