Reitzug Fabian, Luby Stephen P, Pullabhotla Hemant K, Geldsetzer Pascal
Big Data Institute, Nuffield Department of Population Health, University of Oxford, Oxford, United Kingdom.
Woods Institute for the Environment, Stanford University, Palo Alto, CA, United States.
JMIR Res Protoc. 2022 Aug 10;11(8):e35249. doi: 10.2196/35249.
Determining the longer-term health effects of air pollution has been difficult owing to the multitude of potential confounding variables in the relationship between air pollution and health. Air pollution in many areas of South Asia is seasonal, with large spikes in particulate matter (PM) concentration occurring in the winter months. This study exploits this seasonal variation in PM concentration through a natural experiment.
This project aims to determine the causal effect of PM exposure during pregnancy on pregnancy and child health outcomes.
We will use an instrumental variable (IV) design whereby the estimated month of conception is our instrument for exposure to PM with a diameter less than 2.5 μm (PM2.5) during pregnancy. We will assess the plausibility of our assumption that timing of conception is exogenous with regard to our outcomes of interest and will adjust for date of monsoon onset to control for confounding variables related to harvest timing. Our outcomes are 1) birth weight, 2) pregnancy termination resulting in miscarriage, abortion, or still birth, 3) neonatal death, 4) infant death, and 5) child death. We will use data from the Demographic and Health Surveys (DHS) conducted in relevant regions of Bangladesh, India, Nepal, and Pakistan, along with monthly gridded data on PM2.5 concentration (0.1°×0.1° spatial resolution), precipitation data (0.5°×0.5° resolution), temperature data (0.5°×0.5°), and agricultural land use data (0.1°×0.1° resolution).
Data access to relevant DHSs was granted on June 6, 2021 for India, Nepal, Bangladesh, August 24, 2021 for Pakistan, and June 19 2022 for the latest DHS from India.
If the assumptions for a causal interpretation of our instrumental variable analysis are met, this analysis will provide important causal evidence on the maternal and child health effects of PM2.5 exposure during pregnancy. This evidence is important to inform personal behavior and interventions, such as the adoption of indoor air filtration during pregnancy as well as environmental and health policy.
INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/35249.
由于空气污染与健康之间的关系存在众多潜在的混杂变量，确定空气污染对健康的长期影响一直很困难。南亚许多地区的空气污染具有季节性，冬季颗粒物（PM）浓度会大幅飙升。本研究通过自然实验利用了PM浓度的这种季节性变化。
本项目旨在确定孕期暴露于PM对妊娠和儿童健康结局的因果效应。
我们将采用工具变量（IV）设计，即估计的受孕月份作为孕期暴露于直径小于2.5μm的PM（PM2.5）的工具变量。我们将评估受孕时间相对于我们感兴趣的结局是外生的这一假设的合理性，并将调整季风开始日期以控制与收获时间相关的混杂变量。我们的结局包括：1）出生体重；2）导致流产、堕胎或死产的妊娠终止；3）新生儿死亡；4）婴儿死亡；5）儿童死亡。我们将使用在孟加拉国、印度、尼泊尔和巴基斯坦相关地区进行的人口与健康调查（DHS）的数据，以及关于PM2.5浓度（0.1°×0.1°空间分辨率）、降水数据（0.5°×0.5°分辨率）、温度数据（0.5°×0.5°）和农业土地利用数据（0.1°×0.1°分辨率）的月度网格数据。
2021年6月6日获得了印度、尼泊尔、孟加拉国相关DHS的数据访问权限，2021年8月24日获得了巴基斯坦的数据访问权限，2022年6月19日获得了印度最新DHS的数据访问权限。
如果满足我们工具变量分析因果解释的假设，该分析将为孕期暴露于PM2.5对母婴健康的影响提供重要的因果证据。这一证据对于指导个人行为和干预措施（如孕期采用室内空气过滤）以及环境与健康政策具有重要意义。
国际注册报告识别码（IRRID）：DERR1-10.2196/35249
