期刊名称：International Journal of Advances in Engineering and Management
电子版ISSN：2395-5252
出版年度：2022
卷号：4
期号：1
页码：565-571
DOI：10.35629/5252-0401333348
语种：English
出版社：IJAEM JOURNAL
摘要：This research is based on the development of a multivariate logistic regression model to assess the effect of various risk factors such as age, cholesterol level, smoking, alcohol consumption, gender, marital status, occupation status, hypertension, and diabetes family history on the prevalence of diabetes in Gusau local government of Zamfara State. The calculated odds ratio of hypertension 2.608 (260.8%), alcohol consumption 2.405 (240.5%), smoking 2.261 (226.1%), high cholesterol level 2.240 (224.0%), family history of diabetes 2.153 (215.3%), and age 1.019 (101.9%) showed that people with these underlying conditions have a higher risk of developing diabetes than those who do not. A multiple logistic regression model was built with all of the predictor variables, and the coefficients were tested for significance. On the one hand, the Wald test demonstrates that age, cholesterol level, smoking, alcohol intake, hypertension, and a family history of diabetes are significant predictors. Gender, occupation, and marital status, on the other hand, are not statistically significant predictors.Model selection criterion such as the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC) and Hosmer-Lemeshow goodness of fit test were all employed and hence the model with only significant parameters outperformed the model with all parameters. As a result, the proposed logistic regression model can be an effective statistical tool for identifying the relationship between the most significant risk factors to diabetes and thus promptly alerting to take necessary actions to lower the chance of developing the disease.
