An empirical model to predict arsenic pollution affected life expectancy

Abstract

A robust, globally implementable and simple empirical model to predict the arsenic pollution affected life expectancy using a stepwise regression was developed. Life expectancy calculated using a life table technique requires crude death rates data that are not available for small administrative units, complex calculations and does not consider socioeconomic parameters. Hence, a model was needed to forecast the impact of arsenic pollution and socioeconomic parameters on life expectancy for locations with limited data availability. A linear multiple regression technique was used to develop an empirical model to predict arsenic pollution affected life expectancy at birth. The model was calibrated using nine arsenic polluted administrative blocks of district Murshidabad, West Bengal, India and tested independently for three other arsenic polluted blocks of the same district. The R 2 values for the plot of actual versus predicted life expectancy at birth were 0.98 for calibration, testing and independent validation. The model is complementary to the life table technique and offers a means to assist planning by public health engineers and health policy makers to mitigate arsenic pollution on a community priority basis.