Mortality and Air Quality Regulation in Two Indian Cities

描述（由申请人提供）：空气质量干预在发展中国家是一个新现象。印度首都德里近年来采取了一系列空气质量干预措施，为评估这些干预措施对空气质量变化的健康影响提供了一个自然的实验环境。本应用程序的两个总体目标是：(a)评估卫星遥感在不同地理尺度上估计每日空气质量的可行性，特别是对于缺乏空气污染监测覆盖的发展中国家；(b)通过区分管制前（2000- 2002年）和管制后（2004- 2006年）时期的空气污染，评估在德里采取这些干预措施后减轻的死亡率负担。该研究区域没有详细的空气污染数据。然而，2003年7月至12月在德里及其周边113个站点收集的空气污染现场活动数据将用于开发这些站点的地面测量与卫星气溶胶光学深度（AOD）之间的经验关系。利用这种经验关系，每日AOD将用于预测空气动力学直径=2.5¿m, = 10¿m（分别为PM2.5和PM10）和PM10-2.5的每日估计，在控制气象条件的空间分辨率= 5km后。虽然可以使用来自各种传感器的数据来计算AOD，但我们将使用Terra和Aqua卫星上的中分辨率放射光谱仪（MODIS）的数据，因为这两颗卫星一起捕获峰值和非峰值的空气质量估计。由于云层覆盖而丢失的卫星数据将在贝叶斯框架内使用多种插值技术进行填充。对2004年1 - 4月访问的4000名被试进行个人暴露与环境暴露的实证关系研究。反过来，这种关系将用于预测整个研究期间所有死亡病例的个人暴露。贝叶斯层次模型将在不同时空尺度下，参照估计的个人暴露和潜在混杂因素，来检验特定原因的道德。为了评估空气质量干预对道德的影响，我们提出了不同时间（即监管前后时间窗口）对婴儿、具有相对稳定迁移模式的年龄组（30至40岁）和不同年龄-性别群体以及德里和坎普尔在监管前后期间的死亡率和个人暴露关联的差异。公共卫生相关性：这项研究的结果将增进我们对发展中国家空气质量干预措施对健康的好处的理解，这反过来将有助于在发展中国家的其他城市执行类似的条例，以保护人类健康。

DESCRIPTION (provided by applicant): Air quality intervention in developing countries is a new phenomenon. Delhi, the Capital of India, has witnessed a series of air quality interventions in recent years, which provides a natural experimental setting to evaluate the health effects of change in air quality in response to these interventions. Two overarching goals of this application are: (a) to evaluate the feasibility of satellite remote sensing to estimate daily air quality at different geographic scales, particularly for developing countries, which lack adequate coverage of air pollution monitoring, and (b) to assess the burden of mortality alleviated in response to these interventions in Delhi by differencing the cause-specific mortality with reference to air pollution in pre- (2000-02) and post-regulation (2004-06) periods. The spatially detailed air pollution data are unavailable for the study area. Nevertheless, an air pollution field campaign data collected during July-December 2003 at 113 sites in Delhi and its surroundings will be used to develop an empirical relation between ground measurements at these sites and satellite based aerosol optical depth (AOD). Exploiting this empirical relationship, the daily AOD will be used to predict daily estimates of airborne particles =2.5¿m, = 10¿m in aerodynamic diameter (PM2.5 and PM10, respectively) and PM10-2.5, after controlling for meteorological conditions at a spatial resolution = 5km. Although AOD can be computed using the data from various sensors, we will use data from Moderate Resolution Radiospectrometer (MODIS), onboard Terra and Aqua satellites, because these two satellites together capture both peak and off-peak estimates of air quality. Satellite data missing due to cloud cover will be filled using multiple imputation techniques within a Bayesian framework. An empirical relationship between the personal exposure of 4000 subjects (interviewed from January-April 2004) and their ambient exposure will be established. This relationship, in turn, will be used to predict the personal exposure for all cases of mortality for the entire study period. A Bayesian hierarchical model will be employed to examine cause- specific morality with reference to estimated personal exposure and potential confounders at different spatial-temporal scales. To evaluate the impact of air quality interventions on morality we propose differencing mortality and personal exposure association for infants, age cohorts with relatively stable migration pattern (30 to 40 years) and different age-sex groups across time, i.e. pre and post regulation time windows, and across Delhi and Kanpur in both pre- and post-regulation periods. PUBLIC HEALTH RELEVANCE: The findings of this research will advance our understanding of the health benefits of air quality interventions in developing countries, which in turn will be useful for enforcing the similar regulations in other cities of developing countries to protect human health.