Orru Hans, Teinemaa Erik, Lai Taavi, Tamm Tanel, Kaasik Marko, Kimmel Veljo, Kangur Kati, Merisalu Eda, Forsberg Bertil
Department of Public Health, University of Tartu, Tartu 50411, Estonia.
Environ Health. 2009 Mar 3;8:7. doi: 10.1186/1476-069X-8-7.
Health impact assessments (HIA) use information on exposure, baseline mortality/morbidity and exposure-response functions from epidemiological studies in order to quantify the health impacts of existing situations and/or alternative scenarios. The aim of this study was to improve HIA methods for air pollution studies in situations where exposures can be estimated using GIS with high spatial resolution and dispersion modeling approaches.
Tallinn was divided into 84 sections according to neighborhoods, with a total population of approx. 390,000 persons. Actual baseline rates for total mortality and hospitalization with cardiovascular and respiratory diagnosis were identified. The exposure to fine particles (PM2.5) from local emissions was defined as the modeled annual levels. The model validation and morbidity assessment were based on 2006 PM10 or PM2.5 levels at 3 monitoring stations. The exposure-response coefficients used were for total mortality 6.2% (95% CI 1.6-11%) per 10 microg/m3 increase of annual mean PM2.5 concentration and for the assessment of respiratory and cardiovascular hospitalizations 1.14% (95% CI 0.62-1.67%) and 0.73% (95% CI 0.47-0.93%) per 10 microg/m3 increase of PM10. The direct costs related to morbidity were calculated according to hospital treatment expenses in 2005 and the cost of premature deaths using the concept of Value of Life Year (VOLY).
The annual population-weighted-modeled exposure to locally emitted PM2.5 in Tallinn was 11.6 microg/m3. Our analysis showed that it corresponds to 296 (95% CI 76528) premature deaths resulting in 3859 (95% CI 10236636) Years of Life Lost (YLL) per year. The average decrease in life-expectancy at birth per resident of Tallinn was estimated to be 0.64 (95% CI 0.17-1.10) years. While in the polluted city centre this may reach 1.17 years, in the least polluted neighborhoods it remains between 0.1 and 0.3 years. When dividing the YLL by the number of premature deaths, the decrease in life expectancy among the actual cases is around 13 years. As for the morbidity, the short-term effects of air pollution were estimated to result in an additional 71 (95% CI 43-104) respiratory and 204 (95% CI 131-260) cardiovascular hospitalizations per year. The biggest external costs are related to the long-term effects on mortality: this is on average euro 150 (95% CI 40-260) million annually. In comparison, the costs of short-term air-pollution driven hospitalizations are small euro 0.3 (95% CI 0.2-0.4) million.
Sectioning the city for analysis and using GIS systems can help to improve the accuracy of air pollution health impact estimations, especially in study areas with poor air pollution monitoring data but available dispersion models.
健康影响评估（HIA）利用来自流行病学研究的暴露信息、基线死亡率/发病率以及暴露-反应函数，以量化现有情况和/或替代情景对健康的影响。本研究的目的是改进在可使用高空间分辨率地理信息系统（GIS）和扩散模型方法估算暴露的情况下，用于空气污染研究的HIA方法。
根据街区将塔林划分为84个区域，总人口约39万。确定了全因死亡率以及心血管和呼吸系统诊断的住院率的实际基线率。将本地排放的细颗粒物（PM2.5）暴露定义为模拟的年度水平。模型验证和发病率评估基于3个监测站2006年的PM10或PM2.5水平。所使用的暴露-反应系数为：年度平均PM2.5浓度每增加10微克/立方米，全因死亡率为6.2%（95%置信区间1.6 - 11%）；PM10每增加10微克/立方米，呼吸系统和心血管系统住院率评估分别为1.14%（95%置信区间0.62 - 1.67%）和0.73%（95%置信区间0.47 - 0.93%）。根据2005年的医院治疗费用以及使用生命年价值（VOLY）概念计算过早死亡的成本，得出与发病率相关的直接成本。
塔林本地排放的PM2.5的年度人口加权模拟暴露为11.6微克/立方米。我们的分析表明，这相当于每年296例（95%置信区间76 - 528）过早死亡，导致3859例（95%置信区间1023 - 6636）生命年损失（YLL）。塔林每位居民出生时预期寿命的平均减少估计为0.64年（95%置信区间0.17 - 1.10）。在污染严重的市中心，这一数字可能达到1.17年，而在污染最轻的街区则保持在0.1至0.3年之间。将YLL除以过早死亡人数，实际病例中的预期寿命减少约为13年。至于发病率，估计空气污染的短期影响每年导致额外71例（95%置信区间43 - 104）呼吸系统住院和204例（95%置信区间131 - 260）心血管系统住院。最大的外部成本与对死亡率的长期影响相关：平均每年为1.5亿欧元（95%置信区间4000万 - 2.6亿欧元）。相比之下，短期空气污染导致的住院成本较小，为30万欧元（95%置信区间20万 - 40万欧元）。
将城市划分区域进行分析并使用GIS系统有助于提高空气污染健康影响估计的准确性，特别是在空气污染监测数据较差但有可用扩散模型的研究区域。
