Cloudy with a Chance of Cardiovascular Disease: Spatial Interpolation of Mortality Attributable to Air Pollution in China

Abstract

Ambient air pollution, especially particulate matter smaller than 2.5 microns (PM2.5), is a significant contributor to premature mortality. This issue is particularly relevant in China due to rapid urbanization and population growth over the past several decades. Increased pollution in major population centers has resulted in massive loss of life via respiratory and cardiovascular disease. For the Global Burden of Disease (GBD) 2013 Risk Factors analysis, grid-level exposure estimates of annual PM2.5 and cause-specific integrated exposure response (IER) curves were used to predict the relative risk of mortality from a related cause for every pixel. These gridded relative risks were combined with cause-specific provincial mortality results in order to quantify the mortality attributable to air pollution for every province. To provide increased spatial granularity that can further inform the literature and policy-makers, cause-specific mortality was interpolated in order to generate predictions at the pixel level. This analysis demonstrated spatial and temporal trends in mortality attributable to air pollution in China over the GBD time series from 1990 to 2013. Mortality rates were shown to rise over time as a function of PM2.5 exposure, while total mortality increases were most prominent in several key population centers. The cause composition of total attributable mortality was found to have changed over the time series as well, with background rates of LRI and COPD decreasing so that cardiovascular disease makes up the bulk of mortality. As demonstrated by the non-linear concave shape of the IER curve, the distribution of mortality across levels of PM2.5 indicates that the majority of deaths are occurring in areas with moderate levels of air pollution and high population. Future interventions must focus on generating significant decreases in PM2.5 exposure across the Eastern seaboard, where Chinese urbanization is concentrated. This work is limited by the assumption of constant underlying mortality rates within provinces. As the spatial granularity of GBD mortality estimates continues to improve, these estimates will become more accurate and have increased utility to environmental health policy.