Extreme heat events and pregnancy duration: a national study

项目总结/摘要 最近的研究表明，高环境温度会增加早产的风险（<37周）。 妊娠），这是婴儿死亡和长期神经系统残疾的主要原因。早产儿 （37-38周）也有更多的发病率相比，足月分娩。根据气候预测，热浪 预计将增加频率，强度和持续时间，其中许多将导致环境空气质量增加 污染物浓度。拟议的研究旨在利用现有的大型数据库和强大的 在多个空间尺度的方法，以测试总体假设，极端高温 事件会增加早产和早产的风险，假设更强的相关性是 在持续时间更长、强度更大的高温事件之后观察到。使用国家出生记录数据， 国家卫生统计中心，我们将评估美国114个大城市（包括 54%的人口）在县级空间分辨率超过36年的时间（1981-2016）。我们将 此外，还可以从美国人口众多、地理位置具有代表性的八个州获得出生记录数据 （覆盖40%的人口），以评估邮政编码或更细的关系 决议，并检查可能调解的热浪协会伴随着空气中的变化 污染水平。气象学的特点是综合多个气象站每小时的数据 网络和卫星资源，利用每个数据集的优势， 覆盖范围和最小化曝光预测误差。12种污染物的8个环境浓度 选定的州还将通过结合社区多尺度空气质量模型（CMAQ） 带监控测量的输出。早产和早产的统计模型将解释 怀孕的季节性模式（以前研究中可能的偏差来源），以及使用 贝叶斯分层模型将用于联合收割机整合研究地点的信息，并评估 气候区域的异质性，热事件的时间（季节内或数十年），母体 特征（教育程度、种族/民族）和特定于位置的属性（例如，上下文 社会经济指标、空调普及率、城市化）。这种异质性的精确估计是 由于样本特别大，这也允许检查使用 强度和持续时间阈值高于先前评估的阈值。多尺度方法有助于 评估和传播由于暴露预测误差、空间聚集和 怀孕期间的居住流动性。结果可用于通知当地公共卫生预警系统， 作为针对孕妇的热疗法，其最终目标是减少早产及其后遗症。 这项研究还将通过建立一个气候和健康的未来研究产生持久的利益， 美国大陆的综合、空间和时间分辨的、公开可用的气象数据集。

PROJECT SUMMARY/ABSTRACT Recent studies suggest high ambient temperatures increase the risk of preterm birth (<37 completed weeks of gestation), a leading cause of infant mortality and long-term neurological disabilities. Infants born early term (37-38 weeks) also have more morbidity compared to full term births. Under climate projections, heat waves are expected to increase in frequency, intensity, and duration, and many will cause increases in ambient air pollutant concentrations. The proposed research seeks to use large existing databases and robust methodological approaches at multiple spatial scales to test the overarching hypothesis that extreme heat events increase the risk of preterm birth and early term birth, with stronger associations hypothesized to be observed following heat events of longer duration and greater intensity. Using national birth record data from the National Center for Health Statistics, we will assess these relationships in 114 large U.S. cities (covering 54% of the population) at a county-level spatial resolution over a 36-year period (1981-2016). We will additionally obtain birth record data from eight populous and geographically representative U.S. states (covering 40% of the population) over the period 1990-2016 to assess relationships at ZIP code or finer resolution and to examine possible mediation of heat wave associations by accompanying changes in air pollution levels. Meteorology will be characterized by integrating hourly data from multiple weather station networks and satellite-resources, harnessing the strengths of each dataset to maximize spatial and temporal coverage and minimize exposure prediction error. Ambient concentrations of 12 pollutants for the eight selected states will be also characterized by combining Community Multiscale Air Quality Model (CMAQ) outputs with monitor measurements. The statistical models for preterm and early term birth will account for seasonal patterns of conception (a possible source of bias in previous studies), and two-stage analyses using Bayesian hierarchical models will be used to combine information across study locations and assess heterogeneity by climate region, timing of the heat event (within season or across decades), maternal characteristics (educational attainment, race/ethnicity), and location-specific attributes (e.g., contextual socioeconomic indicators, air conditioning prevalence, urbanicity). Precise estimation of this heterogeneity is possible due to the exceptionally large sample, which also allows for examination of heat events defined using higher intensity and duration thresholds than previously assessed. The multi-scale approach facilitates assessment and propagation of uncertainty due to exposure prediction errors, spatial aggregation, and residential mobility during pregnancy. Results can be used to inform local public health warning systems such as heat advisories that target pregnant women with the ultimate goal of reducing early birth and its sequelae. The study will also yield lasting benefits for future studies of climate and health through the creation of an integrated, spatial and temporally-resolved, publically-available meteorology dataset for the continental U.S.