School of Public Health, Key Lab of Public Health Safety of the Ministry of Education and Key Lab of Health Technology Assessment of the Ministry of Health, Fudan University, Shanghai 200032, China.
School of Public Health, Key Lab of Public Health Safety of the Ministry of Education and Key Lab of Health Technology Assessment of the Ministry of Health, Fudan University, Shanghai 200032, China; Shanghai Key Laboratory of Meteorology and Health, Shanghai Meteorological Service, Shanghai, China.
Environ Int. 2018 Sep;118:194-202. doi: 10.1016/j.envint.2018.05.050. Epub 2018 Jun 6.
Epidemiologic studies of PM (particulate matter with aerodynamic diameter ≤2.5 μm) and black carbon (BC) typically use ambient measurements as exposure proxies given that individual measurement is infeasible among large populations. Failure to account for variation in exposure will bias epidemiologic study results. The ability of ambient measurement as a proxy of exposure in regions with heavy pollution is untested.
We aimed to investigate effects of potential determinants and to estimate PM and BC exposure by a modeling approach.
We collected 417 24 h personal PM and 130 72 h personal BC measurements from a panel of 36 nonsmoking college students in Shanghai, China. Each participant underwent 4 rounds of three consecutive 24-h sampling sessions through December 2014 to July 2015. We applied backwards regression to construct mixed effect models incorporating all accessible variables of ambient pollution, climate and time-location information for exposure prediction. All models were evaluated by marginal R and root mean square error (RMSE) from a leave-one-out-cross-validation (LOOCV) and a 10-fold cross-validation (10-fold CV).
Personal PM was 47.6% lower than ambient level, with mean (±Standard Deviation, SD) level of 39.9 (±32.1) μg/m; whereas personal BC (6.1 (±2.8) μg/m) was about one-fold higher than the corresponding ambient concentrations. Ambient levels were the most significant determinants of PM and BC exposure. Meteorological and season indicators were also important predictors. Our final models predicted 75% of the variance in 24 h personal PM and 72 h personal BC. LOOCV analysis showed an R (RMSE) of 0.73 (0.40) for PM and 0.66 (0.27) for BC. Ten-fold CV analysis showed a R (RMSE) of 0.73 (0.41) for PM and 0.68 (0.26) for BC.
We used readily accessible data and established intuitive models that can predict PM and BC exposure. This modeling approach can be a feasible solution for PM exposure estimation in epidemiological studies.
由于个体测量在大人群中不可行，PM（空气动力学直径≤2.5μm 的颗粒物）和黑碳（BC）的流行病学研究通常使用环境测量作为暴露替代物。未能考虑暴露的变化会使流行病学研究结果产生偏差。在污染严重的地区，环境测量作为暴露替代物的能力尚未得到验证。
我们旨在通过建模方法研究潜在决定因素的影响，并估计 PM 和 BC 的暴露量。
我们从中国上海的 36 名不吸烟大学生的小组中收集了 417 个 24 小时个人 PM 和 130 个 72 小时个人 BC 测量值。每个参与者都通过 2014 年 12 月至 2015 年 7 月期间的四轮连续 24 小时采样来进行 3 次连续的 24 小时采样。我们应用向后回归来构建混合效应模型，该模型包含所有可访问的环境污染物、气候和时间-位置信息，用于暴露预测。通过从留一交叉验证（LOOCV）和 10 倍交叉验证（10 倍 CV）中获得的边缘 R 和均方根误差（RMSE）来评估所有模型。
个人 PM 比环境水平低 47.6%，平均（±标准差，SD）水平为 39.9（±32.1）μg/m；而个人 BC（6.1（±2.8）μg/m）则比相应的环境浓度高约一倍。环境水平是 PM 和 BC 暴露的最显著决定因素。气象和季节指标也是重要的预测因子。我们的最终模型预测了 24 小时个人 PM 和 72 小时个人 BC 的 75%的方差。LOOCV 分析表明 PM 的 R（RMSE）为 0.73（0.40），BC 的 R（RMSE）为 0.66（0.27）。10 倍 CV 分析表明 PM 的 R（RMSE）为 0.73（0.41），BC 的 R（RMSE）为 0.68（0.26）。
我们使用了易于获取的数据并建立了直观的模型，可以预测 PM 和 BC 的暴露量。这种建模方法可以成为流行病学研究中 PM 暴露估计的可行解决方案。
Zotero 插件