Identifying the most important life expectancy factors for individual U.S. counties

Abstract

The main goal of this study is to identify the most important life expectancy factors for individual U.S. counties to allow county governments to prioritize the local factors to improve life expectancy. The methodology is to use Geographically Weighted Regression (GWR) to build local regression models for individual counties based on the six most significant life expectancy variables derived from nationwide Ordinary Least Squares (OLS) regression. The results show that Smoking is the most critical life expectancy factor for the majority, approximately 45% of the 3108 contiguous U.S. counties. The second major factor being most important to local counties is Inconsistent Food Supply, for 25% of the total counties. Other four significant nationwide variables, including Insufficient Sleep, Single Parent Household, Physical Inactivity, and Few Fruits and Vegetables, are the most important for 14%–1% of the total counties. Results of this study allow individual counties to target their most influential life expectancy factors to formulate public health strategies and policies that would most effectively improve the county's life expectancy. In addition, findings of the regional patterns of life expectancy factors would help state governments to more efficiently prioritize and allocate health resources in support of county public health programs.