Air pollution has become a major concern worldwide. Many epidemiological studies have proved relationships between fine particulate matter (PM2.5) and various diseases, but most studies only use short-term and models for specific groups to derive relationships with acute diseases. This makes it difficult to understand long-term exposure, nonlinear relationships, and spatial-temporal health risks regarding chronic diseases. Therefore, this study proposed to analyze and map PM2.5 exceedance probability from long-term spatial-temporal monitoring data using radial basis function estimation. We then constructed and compared multiple linear regression and generalized additive models to investigate linear and nonlinear relationships between long-term average PM2.5 concentration, PM2.5 potential probability for exceeding the standard, and standardized mortality for the top ten causes of death in all towns and villages in Taiwan nationally from 2010 to 2017. Linear models indicate that increasing PM2.5 concentration increased malignant neoplasm, pneumonia, and chronic lower respiratory disease mortalities; chronic liver diseases; and cirrhosis; whereas heart diseases and esophagus cancer mortality decreased. For the nonlinear model results, it can be found that there were also significant nonlinear relationships between PM2.5 concentration and malignant mortalities for neoplasm, heart disease, diabetes; and trachea, bronchus, lung, liver, intrahepatic bile duct, and esophagus cancer. Thus, long-term exposure to PM2.5 may be a significant risk factor for multiple acute and chronic diseases. Results from this study can be directly applied worldwide to provide air quality and health management references for governments, and important information on long-term health risks for local residents in the study area.
Keywords: Air pollution; Chronic diseases; Generalized additive model; PM2.5 potential map; Risk analysis; Ten leading causes of death.
© 2021. The Author(s), under exclusive licence to Springer Nature B.V.