Huang Wanyu, Robinson Lucy F, Auchincloss Amy H, Schinasi Leah H, Moore Kari, Melly Steven, Forrest Christopher B, Kenyon Chén C, De Roos Anneclaire J
Department of Epidemiology and Biostatistics, Dornsife School of Public Health, Drexel University, 3215 Market St, Philadelphia, PA, 19104, USA.
Urban Health Collaborative, Dornsife School of Public Health, Drexel University, Philadelphia, PA, USA.
Environ Sci Pollut Res Int. 2025 Feb;32(10):6041-6052. doi: 10.1007/s11356-025-36089-w. Epub 2025 Feb 19.
Childhood asthma exacerbation has multiple risk factors that occur concurrently in the environment - including extreme meteorological conditions, air pollution, aeroallergens, and respiratory virus infections. Few studies have predicted asthma exacerbation based on multiple time-varying environmental risk factors, together. In this study, we constructed an autoregressive integrated moving average (ARIMA) model to predict "high-risk" days for childhood asthma exacerbation in Philadelphia, PA from 2011 to 2016, during the aeroallergen season of each year, using a total of 28,540 asthma exacerbation case events identified from electronic health record (EHR) data. We selected predictors from quantile weighted sum regression (gQWS), incorporating temporal lags and season-stratification (early- vs. late-season), which were entered subsequently into multivariable ARIMA models. We found that daily nitrogen dioxide (NO), as well as monthly rhinovirus and respiratory syncytial virus (RSV) infection levels, were higher on the predicted "high-risk" days, as compared to days with lower childhood asthma exacerbation risk. The model performed better for late-season asthma exacerbation (July to October) than for early season (March to June). Future work and continued research is needed to facilitate local health guidelines pertaining to childhood asthma exacerbation.
儿童哮喘急性发作有多种在环境中同时出现的风险因素，包括极端气象条件、空气污染、气传变应原和呼吸道病毒感染。很少有研究基于多种随时间变化的环境风险因素共同预测哮喘急性发作。在本研究中，我们构建了一个自回归积分滑动平均（ARIMA）模型，利用从电子健康记录（EHR）数据中识别出的总共28540例哮喘急性发作病例事件，预测2011年至2016年宾夕法尼亚州费城每年气传变应原季节儿童哮喘急性发作的“高危”日。我们从分位数加权和回归（gQWS）中选择预测因子，纳入时间滞后和季节分层（季节早期与晚期），随后将其纳入多变量ARIMA模型。我们发现，与儿童哮喘急性发作风险较低的日子相比，预测的“高危”日的每日二氧化氮（NO）以及每月鼻病毒和呼吸道合胞病毒（RSV）感染水平更高。该模型对季节晚期（7月至10月）哮喘急性发作的预测效果比对季节早期（3月至6月）更好。需要开展未来工作并持续进行研究，以推动制定与儿童哮喘急性发作相关的地方健康指南。
