Environmental Data Integration to Assess Impacts of Shale Gas Development on Perinatal Health Outcomes and Childhood Cancers

项目摘要 虽然环境空气污染是一个公认的对婴儿和儿童健康不利的风险因素， 对页岩气等新兴产业的空气排放对健康的影响研究不足 发展（SGD）。SGD行业从2000年的不到30，000个站点迅速扩展到300，000多个站点 到2016年，大约有1760万美国人居住在钻井现场一英里范围内。鉴于这种 大量人群暴露，迫切需要确定与SGD相关的健康风险 空气排放，并制定有效的方法来评估和减少对新出现的危险的接触。本研究 将利用数据科学和大数据技术，将环境数据与健康信息整合， SGD对在子宫内暴露于SGD行业的婴儿和儿童的影响。具体健康数据 将来自一个大型回顾性出生队列（n= 5，275，799），其中包含完整的母亲地址， 出生缺陷和儿童癌症登记从1996年到2009年，这对应于快速增长 德州SGD活动得克萨斯州是美国最大的页岩气生产商，其人口的16%（4.5 100万人）生活在钻井1英里范围内，因此这是研究这种暴露的理想队列。目标1建立小说 来自管理和专有数据源的时空暴露度量，以捕获多种途径 SGD可能影响当地人口的方式，包括特定的SGD工艺（如生产、燃烧）、交通 以及住宅和钻井之间的风向。迄今为止，这些来源尚未用于 大型数据集成项目。通过评估与政策相关的新加坡元风险，这些指标代表了 与流行病学和风险评估中使用的以往暴露评估相比， 这些研究可以应用于SGD以及未来的威胁。目标2将这些时空度量应用于 地理编码的出生队列，以量化特定SGD过程和相关暴露对不良反应的影响。 出生结果、出生缺陷和儿童癌症。这种分析使用了一种独特的因果推理框架 利用跨学科的流行病学，经济学和本体论方法。健康的结果 分析将进一步了解SGD暴露对围产期健康结局的影响， 有助于降低当地社区风险的政策指导方针。这项研究将综合 空间暴露评估方法，先进的环境健康数据科学技术， cabinet推理模型，以产生稳健的风险估计，新加坡元的风险。拟议研究的结果 将通过提供基础知识， SGD暴露对婴儿和儿童健康影响的证据。除了与新加坡元相关的风险之外， 该项目将建立新的方法来评估其他当地环境危害，并帮助弥合多种 流行病学、暴露评估、数据科学和经济学之间的学科差距， 这是一个在公共卫生研究中不常使用的因果推理框架。

PROJECT SUMMARY While ambient air pollution is a well-recognized risk factor for adverse infant and childhood health outcomes, inadequate research exists on the health impacts of air emissions from emerging industries such as shale gas development (SGD). The SGD industry rapidly expanded from under 30,000 sites in 2000 to over 300,000 sites in 2016, so now approximately 17.6 million Americans now live within one mile of a drilling site. Given this substantial population exposure, there is an immediate need to determine the health risks associated with SGD air emissions and develop effective methods to evaluate and reduce exposure to emerging hazards. This study will use data science and big data techniques to integrate environmental data with health information to assess the impact of SGD on infants and children who are exposed to the SGD industry in utero. Specific health data will be derived from a large retrospective birth cohort (n=5,275,799) with full maternal addresses with linkages to birth defect and childhood cancer registries from 1996 through 2009, which corresponds to the rapid increase in Texas SGD activity. Texas is the largest shale gas producer in the country and 16% of its population (4.5 million people) lives within 1 mile of drilling, thus this is the ideal cohort to study this exposure. Aim 1 builds novel spatial-temporal exposure metrics from administrative and proprietary data sources to capture multiple pathways by which SGD may affect local populations, including specific SGD processes (e.g. production, flaring), traffic from the industry, and wind direction between homes and drilling. To date, these sources have not been used in large-scale data integration projects. By assessing policy-relevant SGD exposures, these metrics represent a substantial advancement over previous exposure assessments used in epidemiology and risk assessment studies, which can be applied to SGD as well as future threats. Aim 2 applies these spatial-temporal metrics to the geocoded birth cohort to quantify the impact of specific SGD processes and related exposures on adverse birth outcomes, birth defects, and childhood cancers. This analysis uses a unique causal-inference framework that leverages cross-disciplinary epidemiological, economic, and ontological methods. The results of the health analyses will provide further insights into which SGD exposures influence perinatal health outcomes as well as the policy guidelines that can help reduce risks for local communities. The proposed research will synthesize spatial exposure assessment methods, advance environmental health data science techniques, and develop causal-inference models to produce robust risk estimates for SGD exposures. Findings from the proposed study will provide a better understanding of how SGD is affecting local communities by providing the foundational evidence for the effects of SGD exposure on infant and children’s health. Beyond the risks associated with SGD, this project will establish novel methods to assess other local environmental hazards and help bridge multiple disciplinary gaps among epidemiology, exposure assessment, data science, and economics by demonstrating a causal inference framework not often applied in public health studies.