Using the generalized linear Model (GLM) to model (specific) chronic diseases
Abstract
As per the World Health Organization, chronic diseases rank highly in incidence,
prevalence, and mortality on a global scale with around 60% of the 56.5 million total
global deaths in 2001. Among the most prevalent of the chronic diseases is diabetes.
Diabetes is rapidly increasing in its rates of incidence and mortality. It was ranked as the
seventh leading cause of mortality in 2016, despite the fact that it was not even featured
in the top 10 list in the year 2000. Furthermore, according to the International Diabetes
Federation there are around 552,400 diagnosed cases of diabetes in Kenya. It is evident
fi"om the above that diabetes is a serious and prevalent chronic disease in the countly. As
a result, the aim of this research project is to mod~l and perform a regression analysis on
diabetes using Kenyan data. The modeling process was done using a Generalized Linear
Model. The estimators are expected to display optimal propetties since GLMs use the
MLE method of parameter estimation. Once the model was completed the model diagnosis
displayed a significant R2, suboptimum deviances and a high AIC value. In conclusion,
the most significant ,risk factors of diabetes in the country are high blood pressure, age,
gender, and level of education.