Mortality Risk Factor Ranking for Non-Communicable Diseases Using Multiple Linear Regression

Given non-communicable diseases (NCDs) as the leading cause of death worldwide, it has become critical for health systems to implement and disseminate effective prevention. Risk control in healthcare identifies and manages mortality risk factors to improve patient outcome. Nonetheless, small mortality risk factors in practice often receive unwarranted attention while the large ones are neglected due to lack of risk factor ranking. To address these issues, in this paper, we develop a mortality risk factor ranking (MRFR) scheme to help better mobilize limited healthcare resources. Specifically, MRFR unearths the leading category of NCDs and the most affected region by the leading category, utilizes multiple linear regression to model the relation between mortality rate and the risk factors of the leading category of NCDs, applies backwards selection to identify primary risk factors (PRFs) using statistical significance, and ranks PRFs in terms of partial sum of squares. We have also provided useful guidance to reduce the leading category of mortality on a global scale.