Potential impact fractions(PIF) for joint risk factors

Abstract

An increase in the number of deaths due to various diseases such as Cardiovascular Diseases as well as Type II Diabetes have been of concern to many governments and individuals. However, due to a lack of research on the combined effects of five main harmful products: sugar, salt, fat, alcohol and smoking, it is difficult for one to understand how the different combinations of intake of these harmful products would increase our risk of diseases. As such, our project uses Copulas to help us have a better understanding of the combined effects of these harmful products so as to reduce our risk of succumbing to these diseases. We will be using the main copulas families: Gaussian, t-Copula, Archimedean copulas (Clayton,Gumbel and Frank copulas) as well as using functions in the R programming software to select the best copula family based on maximum likelihood estimation. Using hypothetical intervention values, we then compare the BIC of these copulas for each age group to obtain the best copula family for each age group.