Chowdhury Arman Hossain, Rahman Md Siddikur
Department of Statistics, Begum Rokeya University, Rangpur, Bangladesh.
PLoS Negl Trop Dis. 2025 Jan 16;19(1):e0012800. doi: 10.1371/journal.pntd.0012800. eCollection 2025 Jan.
Bangladesh is facing a formidable challenge in mitigating waterborne diseases risk exacerbated by climate change. However, a comprehensive understanding of the spatio-temporal dynamics of these diseases at the district level remains elusive. Therefore, this study aimed to fill this gap by investigating the spatio-temporal pattern and identifying the best tree-based ML models for determining the meteorological factors associated with waterborne diseases in Bangladesh.
This study used district-level reported cases of waterborne diseases (cholera, amoebiasis, typhoid and hepatitis A) obtained from the Bangladesh Bureau of Statistics (BBS) and meteorological data (temperature, relative humidity, wind speed, and precipitation) sourced from NASA for the period spanning 2017 to 2020. Exploratory spatial analysis, spatial regression and tree-based machine learning models were utilized to analyze the data.
From 2017 and 2020, Bangladesh reported 73, 606 cholera, 38, 472 typhoid, 2, 510 hepatitis A and 1, 643 amoebiasis disease cases. Among the waterborne diseases cholera showed higher incidence rates in Chapai-Nawabganj (456.23), Brahmanbaria (417.44), Faridpur (225.07), Nilphamari (188.62) and Pirojpur (171.62) districts. The spatial regression model identified mean temperature (β = 12.16, s.e: 3.91) as the significant risk factor of waterborne diseases. The optimal XGBoost model highlighted mean and minimum temperature, relative humidity and precipitation as determinants associated with waterborne diseases in Bangladesh from 2017 to 2020.
The findings from the study, incorporating the One Health perspective, provide insights for planning early warning, prevention, and control strategies to combat waterborne diseases in Bangladesh and similar endemic countries. Precautionary measures and intensified surveillance need to be implemented in certain high-risk districts for waterborne diseases across the country.
孟加拉国在减轻气候变化加剧的水源性疾病风险方面面临着严峻挑战。然而，对这些疾病在地区层面的时空动态的全面了解仍然难以捉摸。因此，本研究旨在通过调查时空模式并确定用于确定孟加拉国与水源性疾病相关的气象因素的最佳基于树的机器学习模型来填补这一空白。
本研究使用了从孟加拉国统计局（BBS）获得的地区层面报告的水源性疾病（霍乱、阿米巴病、伤寒和甲型肝炎）病例以及2017年至2020年期间从美国国家航空航天局（NASA）获取的气象数据（温度、相对湿度、风速和降水）。利用探索性空间分析、空间回归和基于树的机器学习模型对数据进行分析。
2017年至2020年期间，孟加拉国报告了73606例霍乱、38472例伤寒、2510例甲型肝炎和1643例阿米巴病病例。在水源性疾病中，霍乱在查派-纳瓦布甘杰（456.23）、布拉曼巴里亚（417.44）、法里德布尔（225.07）、尼尔帕马里（188.62）和皮罗杰布尔（171.62）等地区的发病率较高。空间回归模型确定平均温度（β = 12.16，标准误：3.91）是水源性疾病的重要风险因素。最优的XGBoost模型强调平均温度和最低温度、相对湿度和降水是2017年至2020年期间孟加拉国与水源性疾病相关的决定因素。
本研究的结果结合“同一健康”视角，为规划孟加拉国及类似流行国家应对水源性疾病的预警、预防和控制策略提供了见解。需要在全国某些水源性疾病高风险地区实施预防措施并加强监测。
