You have no items in your shopping cart.
ABSTRACT
Niger is one of the countries in the world where diabetes prevalence is high. Despite several programmes put in place by government and international agencies to reduce or eliminate diabetes the disease seems to be increasing among Nigerians. This study was carried out to identify the significant predictors of diabetes among various attributes available in the literature.
The predictors variables considered in this study are: age, gender, occupation status, alcohol intake, smoking, cholesterol level, hypertension, and family history of diabetes while the dependent variable was risk of diabetes. This study utilized a three – year period data from 2016 to 2018 obtained from a Federal Medical Centre, Abeokuta in Ogun State. Descriptive statistics were obtained for all variables. Chi – square test and two multivariate logistic regressions (full and reduced) models were fitted and compared based on their predictive ability using the Receiver Operating Characteristics (ROC) curve.
The reduced model with area under the ROC curve of 0.92% and predictive accuracy of 79.4% fits better the data and better predicts the risk of developing diabetes than the full model with area under the ROC curve of 0.82 and predictive accuracy of 65.9%. The fitted logistic regression model can be used along with other medical devices to effectively diagnose the risk of diabetes with the aim of reducing the ailment.
