Global Epidemiology最新文献_第6页

Identifying risk factors of post–COVID-19 conditions with machine learning and deep learning algorithms 利用机器学习和深度学习算法识别covid -19后疾病的风险因素

Global Epidemiology

Pub Date : 2025-09-27 DOI: 10.1016/j.gloepi.2025.100221

Guohai Zhou , Scott P. Kelly , Ling Li , Rongjun Shen , Stephen E. Schachterle , Mitchell Henschel , Leo J. Russo , Xiaofeng Zhou

Introduction

Post–COVID-19 conditions (PCC) affect millions of people in the United States. Early diagnosis and PCC management requires an understanding of the epidemiology and drivers behind PCC in the real world.

Methods

We applied multiple machine learning and deep learning models to a large electronic health database of patients with a recent COVID-19 infection in the United States from 2020 to 2022 to quantitatively evaluate progression to newly developed PCC and identify the individual-level risk factors for developing new PCC at 60, 74, 90, and 120 days following initial SARS-CoV-2 infection.

Results

Patients with newly developed primary or secondary PCC were older; had higher Charleson comorbidity scores; and were more likely to smoke, have a body mass index ≥30, or have hyperlipidemia or hypertension than those without evidence of newly developed PCC. Three different machine learning models used to evaluate both the full study period and the Omicron era (beginning January 2022) consistently identified age, the Charlson comorbidity score, and healthcare utilization within 30 days of the index COVID-19 infection as the leading risk factors for developing new primary or secondary PCC. The presence of disseminated intravascular coagulation at baseline was among the 10 strongest predictors of newly developed cardiovascular or secondary PCC in the full study period and the Omicron era.

Conclusion

Multiple machine learning and deep learning models identified the Charlson comorbidity score, age, and frequency of healthcare utilization, which may help predict the occurrence of new PCC and demonstrated the utility of the models for individualized risk prediction.

在美国，covid -19后疾病（PCC）影响着数百万人。早期诊断和PCC管理需要了解现实世界中PCC背后的流行病学和驱动因素。方法对美国2020年至2022年近期COVID-19感染患者的大型电子健康数据库应用多种机器学习和深度学习模型，定量评估新发PCC的进展情况，并确定在首次感染SARS-CoV-2后60、74、90和120天发生新发PCC的个人层面危险因素。结果新发原发性或继发性PCC患者年龄较大；查理森合并症评分较高；吸烟、体重指数≥30、高脂血症或高血压的可能性高于无新发PCC证据的患者。用于评估整个研究期和Omicron时代（从2022年1月开始）的三种不同的机器学习模型一致将年龄、Charlson合并症评分和COVID-19感染指数30天内的医疗保健利用确定为发展新的原发性或继发性PCC的主要风险因素。在整个研究期间和Omicron时代，基线时弥散性血管内凝血的存在是新发心血管或继发性PCC的10个最强预测因子之一。结论多重机器学习和深度学习模型识别了Charlson合并症评分、年龄和医疗保健使用频率，有助于预测新发PCC的发生，证明了模型在个体化风险预测中的实用性。

{"title":"Identifying risk factors of post–COVID-19 conditions with machine learning and deep learning algorithms","authors":"Guohai Zhou , Scott P. Kelly , Ling Li , Rongjun Shen , Stephen E. Schachterle , Mitchell Henschel , Leo J. Russo , Xiaofeng Zhou","doi":"10.1016/j.gloepi.2025.100221","DOIUrl":"10.1016/j.gloepi.2025.100221","url":null,"abstract":"<div><h3>Introduction</h3><div>Post–COVID-19 conditions (PCC) affect millions of people in the United States. Early diagnosis and PCC management requires an understanding of the epidemiology and drivers behind PCC in the real world.</div></div><div><h3>Methods</h3><div>We applied multiple machine learning and deep learning models to a large electronic health database of patients with a recent COVID-19 infection in the United States from 2020 to 2022 to quantitatively evaluate progression to newly developed PCC and identify the individual-level risk factors for developing new PCC at 60, 74, 90, and 120 days following initial SARS-CoV-2 infection.</div></div><div><h3>Results</h3><div>Patients with newly developed primary or secondary PCC were older; had higher Charleson comorbidity scores; and were more likely to smoke, have a body mass index ≥30, or have hyperlipidemia or hypertension than those without evidence of newly developed PCC. Three different machine learning models used to evaluate both the full study period and the Omicron era (beginning January 2022) consistently identified age, the Charlson comorbidity score, and healthcare utilization within 30 days of the index COVID-19 infection as the leading risk factors for developing new primary or secondary PCC. The presence of disseminated intravascular coagulation at baseline was among the 10 strongest predictors of newly developed cardiovascular or secondary PCC in the full study period and the Omicron era.</div></div><div><h3>Conclusion</h3><div>Multiple machine learning and deep learning models identified the Charlson comorbidity score, age, and frequency of healthcare utilization, which may help predict the occurrence of new PCC and demonstrated the utility of the models for individualized risk prediction.</div></div>","PeriodicalId":36311,"journal":{"name":"Global Epidemiology","volume":"10 ","pages":"Article 100221"},"PeriodicalIF":0.0,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145264871","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Religion as a domain of exposure assessment in epidemiologic studies: History, meaning, and implications 宗教作为流行病学研究中暴露评估的一个领域：历史、意义和影响

Global Epidemiology

Pub Date : 2025-09-26 DOI: 10.1016/j.gloepi.2025.100218

Jeff Levin

For over a century, research findings have accumulated linking measures of religious identity and practice to rates of morbidity and mortality and other population-health outcomes. These studies have been conducted on every continent, have drawn on population samples of nearly all major religions and denominations, and have investigated associations with numerous overall and cause-specific rates of most major chronic diseases and acute conditions, both physical and psychiatric. Yet despite thousands of published studies and many comprehensive reviews, relatively less attention has been paid to conceptual, theoretical, and policy-related issues, notably what these studies are actually assessing, what resultant findings mean and do not mean, and why they should matter. This has contributed, in part, to the continued contentiousness of this subject within some segments of medicine and epidemiology. In an effort to clarify issues, this commentary outlines (a) the current status of exposure assessment related to religion; (b) the ways that the meaning of empirical findings are occasionally misjudged; (c) possible mediators of putative religious influences on morbidity and mortality; and (d) policy implications of existing findings. The latter includes implications for epidemiologic and population-health research, for clinical medicine, for congregational health promotion and disease prevention programs, for global health development and public health policy, and for human flourishing. In sum, this area of research can make a worthwhile contribution, but the burden is on investigators to ensure that their religious measures are validated and that findings are carefully unpacked in terms of their real-world implications for population health.

一个多世纪以来，越来越多的研究结果将宗教认同和宗教实践与发病率和死亡率以及其他人口健康结果联系起来。这些研究是在各大洲进行的，利用了几乎所有主要宗教和教派的人口样本，并调查了大多数主要慢性疾病和急性身体和精神疾病的许多总体和具体原因的发病率之间的关系。然而，尽管发表了数千项研究和许多综合评论，但对概念、理论和政策相关问题的关注相对较少，特别是这些研究实际上在评估什么，最终发现意味着什么，不意味着什么，以及为什么它们应该重要。这在一定程度上导致了这一主题在医学和流行病学的某些领域内持续存在争议。为了澄清问题，本评注概述(a)与宗教有关的暴露评估的现状；(b)实证发现的意义偶尔被误判的方式；(c)假定的宗教对发病率和死亡率的影响的可能中介；(d)现有研究结果的政策含义。后者包括对流行病学和人口健康研究、临床医学、会众健康促进和疾病预防计划、全球健康发展和公共卫生政策以及人类繁荣的影响。总而言之，这一领域的研究可以做出有价值的贡献，但调查人员的责任是确保他们的宗教措施得到验证，并确保研究结果在实际生活中对人口健康的影响得到仔细的解释。

{"title":"Religion as a domain of exposure assessment in epidemiologic studies: History, meaning, and implications","authors":"Jeff Levin","doi":"10.1016/j.gloepi.2025.100218","DOIUrl":"10.1016/j.gloepi.2025.100218","url":null,"abstract":"<div><div>For over a century, research findings have accumulated linking measures of religious identity and practice to rates of morbidity and mortality and other population-health outcomes. These studies have been conducted on every continent, have drawn on population samples of nearly all major religions and denominations, and have investigated associations with numerous overall and cause-specific rates of most major chronic diseases and acute conditions, both physical and psychiatric. Yet despite thousands of published studies and many comprehensive reviews, relatively less attention has been paid to conceptual, theoretical, and policy-related issues, notably what these studies are actually assessing, what resultant findings mean and do not mean, and why they should matter. This has contributed, in part, to the continued contentiousness of this subject within some segments of medicine and epidemiology. In an effort to clarify issues, this commentary outlines (a) the current status of exposure assessment related to religion; (b) the ways that the meaning of empirical findings are occasionally misjudged; (c) possible mediators of putative religious influences on morbidity and mortality; and (d) policy implications of existing findings. The latter includes implications for epidemiologic and population-health research, for clinical medicine, for congregational health promotion and disease prevention programs, for global health development and public health policy, and for human flourishing. In sum, this area of research can make a worthwhile contribution, but the burden is on investigators to ensure that their religious measures are validated and that findings are carefully unpacked in terms of their real-world implications for population health.</div></div>","PeriodicalId":36311,"journal":{"name":"Global Epidemiology","volume":"10 ","pages":"Article 100218"},"PeriodicalIF":0.0,"publicationDate":"2025-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145264454","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Body mass index and mortality in a nationally representative cohort of south African adults 一项具有全国代表性的南非成年人队列的体重指数和死亡率

Global Epidemiology

Pub Date : 2025-09-26 DOI: 10.1016/j.gloepi.2025.100220

Annibale Cois

Aim

To examine the association between Body Mass Index (BMI) and all-cause mortality in South Africa.

Methods

Longitudinal data on adults 20 years and older from five waves (2008, 2010–11, 2012, 2014–15, and 2017) of the South African National Income Dynamics Study were analysed. Survival proportional hazard models, adjusted for sociodemographic and lifestyle characteristics, were used to estimate the relationship between BMI and mortality. Sensitivity analyses were conducted to assess the robustness of the estimates.

Results

Of the 12,402 eligible individuals, 10917 had valid BMI measurements and were included in the analyses. During a total of 83,077 person-years of observation, 1741 individuals died.

Hazard ratios for all-cause mortality were significantly lower in the BMI range 25–40 kg/m² in comparison with the reference category of 18.5–25 kg/m² and were minimal in the range 30–35 kg/m² (HR = 0.68, 95% CI: 0.50–0.88). BMI < 18.5 kg/m² was associated with an increased risk of death, with a maximum hazard ratio of 2.14 (95% CI: 1.36–3.4) in the <16 kg/m² category. The pattern was repeated in the sex-specific analyses. The relationship persisted after restricting the analyses to never smokers, excluding subjects with pre-existing conditions or who died in the first two years of follow-up.

Conclusions

This study suggests that, in the South African adult population, BMI in the overweight or mild obesity range according to international definitions is associated with a reduced risk of mortality compared to the” healthy weight” range. Further research is needed to corroborate these results.

目的探讨南非人身体质量指数（BMI）与全因死亡率之间的关系。方法对南非国民收入动态研究五波（2008年、2010-11年、2012年、2014-15年和2017年）20岁及以上成年人的纵向数据进行分析。经社会人口统计学和生活方式特征调整后的生存比例风险模型用于估计BMI和死亡率之间的关系。进行敏感性分析以评估估计的稳健性。结果在12402名符合条件的个体中，10917名具有有效的BMI测量值，并被纳入分析。在总共83077人年的观察中，1741人死亡。与参考类别18.5-25 kg/m2相比，BMI在25-40 kg/m2范围内的全因死亡率风险比显著降低，而在30-35 kg/m2范围内的全因死亡率风险比最小（HR = 0.68, 95% CI: 0.50-0.88）。BMI 18.5 kg/m2与死亡风险增加相关，在16 kg/m2类别中，最大风险比为2.14 （95% CI: 1.36-3.4）。在性别分析中重复了这种模式。在将分析对象限制为从不吸烟者，排除了已有疾病或在随访前两年死亡的受试者后，这种关系仍然存在。结论：本研究表明，在南非成年人中，根据国际定义，BMI处于超重或轻度肥胖范围内，与“健康体重”范围相比，死亡率风险降低。需要进一步的研究来证实这些结果。

{"title":"Body mass index and mortality in a nationally representative cohort of south African adults","authors":"Annibale Cois","doi":"10.1016/j.gloepi.2025.100220","DOIUrl":"10.1016/j.gloepi.2025.100220","url":null,"abstract":"<div><h3>Aim</h3><div>To examine the association between Body Mass Index (BMI) and all-cause mortality in South Africa.</div></div><div><h3>Methods</h3><div>Longitudinal data on adults 20 years and older from five waves (2008, 2010–11, 2012, 2014–15, and 2017) of the South African National Income Dynamics Study were analysed. Survival proportional hazard models, adjusted for sociodemographic and lifestyle characteristics, were used to estimate the relationship between BMI and mortality. Sensitivity analyses were conducted to assess the robustness of the estimates.</div></div><div><h3>Results</h3><div>Of the 12,402 eligible individuals, 10917 had valid BMI measurements and were included in the analyses. During a total of 83,077 person-years of observation, 1741 individuals died.</div><div>Hazard ratios for all-cause mortality were significantly lower in the BMI range 25–40 kg/m<sup>2</sup> in comparison with the reference category of 18.5–25 kg/m<sup>2</sup> and were minimal in the range 30–35 kg/m<sup>2</sup> (HR = 0.68, 95% CI: 0.50–0.88). BMI < 18.5 kg/m<sup>2</sup> was associated with an increased risk of death, with a maximum hazard ratio of 2.14 (95% CI: 1.36–3.4) in the <16 kg/m<sup>2</sup> category. The pattern was repeated in the sex-specific analyses. The relationship persisted after restricting the analyses to never smokers, excluding subjects with pre-existing conditions or who died in the first two years of follow-up.</div></div><div><h3>Conclusions</h3><div>This study suggests that, in the South African adult population, BMI in the overweight or mild obesity range according to international definitions is associated with a reduced risk of mortality compared to the” healthy weight” range. Further research is needed to corroborate these results.</div></div>","PeriodicalId":36311,"journal":{"name":"Global Epidemiology","volume":"10 ","pages":"Article 100220"},"PeriodicalIF":0.0,"publicationDate":"2025-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145219388","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Mortality trends associated with hypertension and atrial fibrillation: A CDC WONDER data analysis 高血压和房颤相关的死亡率趋势：CDC WONDER数据分析

Global Epidemiology

Pub Date : 2025-09-26 DOI: 10.1016/j.gloepi.2025.100217

Saad Khan , Usama Idrees , Safa Nasir , Fatima Naveed , Aqsa Munir , Muhammad Junaid Iqbal , Rizwan Ahmad , Muhammad Ubaid Hussain , Fathimathul henna , Khansha Saeed , Amin ul Haq , Ali Ahmed

Background

Hypertension (HTN) and atrial fibrillation (AF) are prevalent cardiovascular disorders that significantly increase the risk of serious complications such as stroke and heart failure, leading to elevated mortality rates. Despite the established relationship between HTN and AF, there is a lack of comprehensive evidence on mortality trends and disparities across various demographic groups in the United States.

Objectives

This study aims to analyze the nationwide mortality trends due to HTN-AF from 1999 to 2020 and to identify disparities across different demographics. The goals include understanding the impact of HTN-AF on public health and informing targeted screening and therapeutic strategies.

Methods

Data on mortality figures related to HTN and AF were obtained from the Centers for Disease Control and Prevention's (CDC) Wide-Ranging Online Data for Epidemiologic Research (WONDER) database. The analysis included age-adjusted mortality rates (AAMR) and crude mortality rates (CMR) stratified by gender, age, race/ethnicity, and geographic regions. The Joinpoint software was used to analyze temporal trends in age-adjusted mortality rate (AAMR). Data were obtained from publicly available multiple causes of death records via the CDC WONDER database.

Results

From 1999 to 2020, the AAMR due to hypertension-attributable factors rose sharply from 2.89 to 23.98 per 100,000 (APC: 4.8 %). Males had higher AAMRs than females, with Black or African American populations seeing the steepest increases. Regionally, the West had the highest AAMR, and rural areas experienced the most significant rise, with micropolitan areas showing the highest APC.

Conclusion

HTN-AF mortality has been increasing steadily across all genders, races, and regions. The study underscores the importance of improving healthcare policies, bridging coverage gaps, and enhancing education and awareness to curb these alarming trends. Addressing the disparities in healthcare access and promoting cardiovascular health initiatives are crucial for reducing HTN-AF-related mortality.

背景：高血压（HTN）和心房颤动（AF）是常见的心血管疾病，可显著增加中风和心力衰竭等严重并发症的风险，导致死亡率升高。尽管HTN和房颤之间存在既定的关系，但缺乏关于美国不同人口群体死亡率趋势和差异的综合证据。目的分析1999年至2020年全国HTN-AF死亡率趋势，并确定不同人口统计学差异。目标包括了解HTN-AF对公共卫生的影响，并告知有针对性的筛查和治疗策略。方法HTN和AF相关的死亡率数据来自美国疾病控制与预防中心（CDC）广泛的流行病学研究在线数据（WONDER）数据库。分析包括按性别、年龄、种族/民族和地理区域分层的年龄调整死亡率（AAMR）和粗死亡率（CMR）。使用Joinpoint软件分析年龄调整死亡率（AAMR）的时间趋势。数据来自CDC WONDER数据库中公开的多种死因记录。结果1999 - 2020年，高血压归因因素导致的AAMR由2.89 / 10万急剧上升至23.98 / 10万（APC: 4.8%）。男性的aamr高于女性，其中黑人或非裔美国人的增幅最大。从地区上看，西部地区的AAMR最高，农村地区的APC上升最为显著，而小城市地区的APC最高。结论htn - af死亡率在所有性别、种族和地区均呈稳步上升趋势。该研究强调了改善医疗保健政策、缩小覆盖差距以及加强教育和意识以遏制这些令人担忧的趋势的重要性。解决医疗保健获取方面的差异和促进心血管健康举措对于降低htn - af相关死亡率至关重要。

{"title":"Mortality trends associated with hypertension and atrial fibrillation: A CDC WONDER data analysis","authors":"Saad Khan , Usama Idrees , Safa Nasir , Fatima Naveed , Aqsa Munir , Muhammad Junaid Iqbal , Rizwan Ahmad , Muhammad Ubaid Hussain , Fathimathul henna , Khansha Saeed , Amin ul Haq , Ali Ahmed","doi":"10.1016/j.gloepi.2025.100217","DOIUrl":"10.1016/j.gloepi.2025.100217","url":null,"abstract":"<div><h3>Background</h3><div>Hypertension (HTN) and atrial fibrillation (AF) are prevalent cardiovascular disorders that significantly increase the risk of serious complications such as stroke and heart failure, leading to elevated mortality rates. Despite the established relationship between HTN and AF, there is a lack of comprehensive evidence on mortality trends and disparities across various demographic groups in the United States.</div></div><div><h3>Objectives</h3><div>This study aims to analyze the nationwide mortality trends due to HTN-AF from 1999 to 2020 and to identify disparities across different demographics. The goals include understanding the impact of HTN-AF on public health and informing targeted screening and therapeutic strategies.</div></div><div><h3>Methods</h3><div>Data on mortality figures related to HTN and AF were obtained from the Centers for Disease Control and Prevention's (CDC) Wide-Ranging Online Data for Epidemiologic Research (WONDER) database. The analysis included age-adjusted mortality rates (AAMR) and crude mortality rates (CMR) stratified by gender, age, race/ethnicity, and geographic regions. The Joinpoint software was used to analyze temporal trends in age-adjusted mortality rate (AAMR). Data were obtained from publicly available multiple causes of death records via the CDC WONDER database.</div></div><div><h3>Results</h3><div>From 1999 to 2020, the AAMR due to hypertension-attributable factors rose sharply from 2.89 to 23.98 per 100,000 (APC: 4.8 %). Males had higher AAMRs than females, with Black or African American populations seeing the steepest increases. Regionally, the West had the highest AAMR, and rural areas experienced the most significant rise, with micropolitan areas showing the highest APC.</div></div><div><h3>Conclusion</h3><div>HTN-AF mortality has been increasing steadily across all genders, races, and regions. The study underscores the importance of improving healthcare policies, bridging coverage gaps, and enhancing education and awareness to curb these alarming trends. Addressing the disparities in healthcare access and promoting cardiovascular health initiatives are crucial for reducing HTN-AF-related mortality.</div></div>","PeriodicalId":36311,"journal":{"name":"Global Epidemiology","volume":"10 ","pages":"Article 100217"},"PeriodicalIF":0.0,"publicationDate":"2025-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145264455","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Mathematical modelling and time series clustering of Mpox outbreak: A comparative study of the top 10 affected countries and implications for future outbreak management m痘暴发的数学建模和时间序列聚类：十大受影响国家的比较研究及其对未来暴发管理的影响

Global Epidemiology

Pub Date : 2025-07-27 DOI: 10.1016/j.gloepi.2025.100214

Mark-Daniels Tamakloe , Ametus Kuuwill , Ibrahim Osumanu , Helina Siripi

The 2022 Mpox outbreak, characterized by its rapid cross-continental spread beyond traditionally endemic regions, presented a renewed threat to global health security. This study presents a comparative epidemiological analysis of the ten countries most affected by Mpox, integrating mathematical modelling with time series clustering, the first of its kind to analyze the 2022 WHO Mpox data. By applying an SIR-based model to estimate the effective transmission rate, basic reproduction number, time of first infection, and initial susceptible population, the study captures both the pace and persistence of Mpox spread, while critically assessing the effectiveness of national public health responses. Key findings reveal a paradox in North America: Canada exhibited a high transmission rate but a low reproduction number, indicating an elevated transmission potential per contact alongside limited secondary spread. This is likely due to concurrent containment measures or behavioral factors. In contrast, the United States, despite having a lower initial transmission rate, recorded a higher reproduction number. Similarly, Germany exhibited a similar risk trajectory, with elevated reproductive numbers despite robust infrastructure. The cases in the USA and Germany are likely due to systemic health and socio-political policy gaps and delayed behavior-targeted interventions, particularly in the population of men having sex with men (MSM). In Latin America, countries such as Peru and Mexico suffered disproportionately, likely due to limited access to healthcare, which compounded transmission dynamics and reproductive potential. Our study demonstrates that effective Mpox control is not solely dependent on health infrastructure, but also on behavioral targeting, equity, and adaptive health governance. This calls for cross-country and intercontinental collaborations towards combating current and future health shocks, including epidemics.

2022年麻疹疫情的特点是迅速跨大陆传播，超出了传统流行地区，对全球卫生安全构成了新的威胁。本研究对受麻疹影响最严重的10个国家进行了比较流行病学分析，将数学模型与时间序列聚类相结合，这是首次对2022年世卫组织麻疹数据进行分析。通过应用基于免疫特性的模型来估计有效传播率、基本繁殖数、首次感染时间和初始易感人群，该研究捕捉到了痘传播的速度和持续时间，同时严格评估了国家公共卫生应对措施的有效性。主要发现揭示了北美的一个悖论：加拿大表现出高传播率，但繁殖数量低，表明每次接触的传播潜力升高，同时继发传播有限。这可能是由于同时采取的遏制措施或行为因素造成的。相比之下，尽管美国的初始传播率较低，但其繁殖数量却较高。同样，德国也表现出类似的风险轨迹，尽管基础设施健全，但该国的生育数量仍在上升。美国和德国的病例可能是由于系统性的卫生和社会政治政策差距以及延迟的针对行为的干预，特别是在男男性行为人群中。在拉丁美洲，秘鲁和墨西哥等国家受到的影响尤为严重，可能是由于获得医疗保健的机会有限，这加剧了传播动态和生殖潜力。我们的研究表明，有效的麻疹控制不仅依赖于卫生基础设施，还依赖于行为目标、公平和适应性卫生治理。这就要求开展跨国和洲际合作，以应对当前和未来的健康冲击，包括流行病。

{"title":"Mathematical modelling and time series clustering of Mpox outbreak: A comparative study of the top 10 affected countries and implications for future outbreak management","authors":"Mark-Daniels Tamakloe , Ametus Kuuwill , Ibrahim Osumanu , Helina Siripi","doi":"10.1016/j.gloepi.2025.100214","DOIUrl":"10.1016/j.gloepi.2025.100214","url":null,"abstract":"<div><div>The 2022 Mpox outbreak, characterized by its rapid cross-continental spread beyond traditionally endemic regions, presented a renewed threat to global health security. This study presents a comparative epidemiological analysis of the ten countries most affected by Mpox, integrating mathematical modelling with time series clustering, the first of its kind to analyze the 2022 WHO Mpox data. By applying an SIR-based model to estimate the effective transmission rate, basic reproduction number, time of first infection, and initial susceptible population, the study captures both the pace and persistence of Mpox spread, while critically assessing the effectiveness of national public health responses. Key findings reveal a paradox in North America: Canada exhibited a high transmission rate but a low reproduction number, indicating an elevated transmission potential per contact alongside limited secondary spread. This is likely due to concurrent containment measures or behavioral factors. In contrast, the United States, despite having a lower initial transmission rate, recorded a higher reproduction number. Similarly, Germany exhibited a similar risk trajectory, with elevated reproductive numbers despite robust infrastructure. The cases in the USA and Germany are likely due to systemic health and socio-political policy gaps and delayed behavior-targeted interventions, particularly in the population of men having sex with men (MSM). In Latin America, countries such as Peru and Mexico suffered disproportionately, likely due to limited access to healthcare, which compounded transmission dynamics and reproductive potential. Our study demonstrates that effective Mpox control is not solely dependent on health infrastructure, but also on behavioral targeting, equity, and adaptive health governance. This calls for cross-country and intercontinental collaborations towards combating current and future health shocks, including epidemics.</div></div>","PeriodicalId":36311,"journal":{"name":"Global Epidemiology","volume":"10 ","pages":"Article 100214"},"PeriodicalIF":0.0,"publicationDate":"2025-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144724396","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Statistical inference and effect measures in abstracts of major HIV and AIDS journals, 1987–2022: A systematic review 1987-2022年主要HIV和AIDS期刊摘要的统计推断和效应度量：系统回顾

Global Epidemiology

Pub Date : 2025-07-25 DOI: 10.1016/j.gloepi.2025.100213

Andreas Stang , Henning Schäfer , Ahmad Idrissi-Yaghir , Christoph M. Friedrich , Matthew P. Fox

Objectives

With the emergence of HIV/AIDS journals, the development of the reporting of statistical inference and effect measures in published abstracts can be examined from the beginning in a new field. The aim of this study was to describe time trends of statistical inference and effect measure reporting of major HIV/AIDS journals

Methods

We included 10 major HIV/AIDS journals and analyzed all available PubMed entries for the period 1987 through 2022. We applied rule-based text mining and machine learning methodology to detect the presence of confidence intervals, numerical p-values or comparisons of p-values with thresholds, language describing statistical significance, and effect measures for dichotomous outcomes

Results

Among 41,730 PubMed entries from the major HIV/AIDS journals, 31,665 contained an abstract. In the early years, most abstracts reporting statistical inference contained only significance terminology without confidence intervals and p-values. From 1988 to 2005, each year 30 % of all abstracts contained p-values without confidence intervals. Thereafter, this reporting style continued to decline. The reporting of confidence intervals increased steadily from 1988 (11 %) to 2022 (56 %). Of the 17 % of abstracts in 2017–2022 that included any effect measure, half reported odds ratios (51 %), followed by hazard ratios (28 %) and risk ratios (16 %). Difference measures and number needed to treat or harm were very uncommon

Conclusions

Within the HIV/AIDS literature, there has been widespread use of confidence intervals. Most of the journals that we reviewed had a decrease in reporting only statistical significance without confidence intervals over time

目的随着HIV/AIDS期刊的出现，可以在一个新的领域从头开始考察发表摘要的统计推断和效果度量报告的发展。本研究的目的是描述主要HIV/AIDS期刊的统计推断和效果测量报告的时间趋势。方法我们纳入了10种主要HIV/AIDS期刊，并分析了1987年至2022年期间所有可获得的PubMed条目。我们应用基于规则的文本挖掘和机器学习方法来检测置信区间、数值p值或p值与阈值的比较、描述统计显著性的语言以及二分类结果的效果度量的存在。结果在来自主要HIV/AIDS期刊的41,730篇PubMed条目中，31,665篇包含摘要。在早期，大多数报告统计推断的摘要只包含显著性术语，没有置信区间和p值。从1988年到2005年，每年有30%的摘要包含没有置信区间的p值。此后，这种报道方式继续减少。从1988年（11%）到2022年（56%），置信区间的报告稳步增加。2017-2022年，17%的摘要包含了任何影响测量，其中一半报告了优势比（51%），其次是风险比（28%）和风险比（16%）。治疗或伤害所需的不同措施和数量非常罕见。结论在艾滋病毒/艾滋病文献中，广泛使用置信区间。我们回顾的大多数期刊随着时间的推移，报告的统计意义没有置信区间都有所减少

{"title":"Statistical inference and effect measures in abstracts of major HIV and AIDS journals, 1987–2022: A systematic review","authors":"Andreas Stang , Henning Schäfer , Ahmad Idrissi-Yaghir , Christoph M. Friedrich , Matthew P. Fox","doi":"10.1016/j.gloepi.2025.100213","DOIUrl":"10.1016/j.gloepi.2025.100213","url":null,"abstract":"<div><h3>Objectives</h3><div>With the emergence of HIV/AIDS journals, the development of the reporting of statistical inference and effect measures in published abstracts can be examined from the beginning in a new field. The aim of this study was to describe time trends of statistical inference and effect measure reporting of major HIV/AIDS journals</div></div><div><h3>Methods</h3><div>We included 10 major HIV/AIDS journals and analyzed all available PubMed entries for the period 1987 through 2022. We applied rule-based text mining and machine learning methodology to detect the presence of confidence intervals, numerical <em>p</em>-values or comparisons of p-values with thresholds, language describing statistical significance, and effect measures for dichotomous outcomes</div></div><div><h3>Results</h3><div>Among 41,730 PubMed entries from the major HIV/AIDS journals, 31,665 contained an abstract. In the early years, most abstracts reporting statistical inference contained only significance terminology without confidence intervals and <em>p</em>-values. From 1988 to 2005, each year 30 % of all abstracts contained p-values without confidence intervals. Thereafter, this reporting style continued to decline. The reporting of confidence intervals increased steadily from 1988 (11 %) to 2022 (56 %). Of the 17 % of abstracts in 2017–2022 that included any effect measure, half reported odds ratios (51 %), followed by hazard ratios (28 %) and risk ratios (16 %). Difference measures and number needed to treat or harm were very uncommon</div></div><div><h3>Conclusions</h3><div>Within the HIV/AIDS literature, there has been widespread use of confidence intervals. Most of the journals that we reviewed had a decrease in reporting only statistical significance without confidence intervals over time</div></div>","PeriodicalId":36311,"journal":{"name":"Global Epidemiology","volume":"10 ","pages":"Article 100213"},"PeriodicalIF":0.0,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144748962","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Do certain blood groups increase COVID-19 severity and mortality? 某些血型会增加COVID-19的严重程度和死亡率吗？

Global Epidemiology

Pub Date : 2025-07-08 DOI: 10.1016/j.gloepi.2025.100212

Tegene Atamenta kitaw, Ribka Nigatu Haile

Background

The existence of a relationship between the ABO blood group and COVID-19 severity and mortality is still an unresolved concern. Some studies report that groups O and A show a lower and higher risk of developing severe COVID-19 and mortality, respectively. Some studies also report the reverse. There are inconclusive results from different studies. Thus, this study sought to determine the possible associations of ABO blood type with COVID-19 severity and mortality.

Methods

A retrospective study was conducted among 570 adults with real-time reverse transcription-polymerase chain reaction confirmed positive COVID-19 patients attending Eka Kotebe General Hospital, COVID-19 Treatment Center. A Kaplan-Meier survival curve was computed to examine the difference in survival experience between ABO blood groups. Multinomial and binary logistic regression models were fitted to determine the association between ABO blood group with COVID-19 severity and mortality, respectively.

Result

238 (41.8 %) COVID-19 patients had blood group B, followed by 201 (35.3 %) A, 82 (14.4 %) O, and 49 (8.6 %) AB blood type. 23.68 % of participants develop severe COVID-19. Overall, 15.26 % COVID-19-related mortality was found. No difference in survival experience was observed between ABO blood types. There was no statistically significant association between ABO blood type and COVID-19 severity, and mortality.

Conclusion

We found no relationship between ABO blood group differences and COVID-19 severity, and mortality. Further, well-design-controlled studies are suggested to explore the potential link of ABO blood group with COVID-19 severity and mortality.

ABO血型与COVID-19严重程度和死亡率之间是否存在关系仍是一个未解决的问题。一些研究报告称，O组和A组分别表现出较低和较高的发生严重COVID-19和死亡的风险。一些研究也报告了相反的情况。不同的研究都有不确定的结果。因此，本研究试图确定ABO血型与COVID-19严重程度和死亡率的可能关联。方法回顾性分析在Eka Kotebe总医院COVID-19治疗中心就诊的570例成人实时逆转录聚合酶链反应阳性患者。计算Kaplan-Meier生存曲线来检查ABO血型之间生存经验的差异。拟合多项和二元logistic回归模型，分别确定ABO血型与COVID-19严重程度和死亡率之间的关系。结果B型238例（41.8%），A型201例（35.3%），O型82例（14.4%），AB型49例（8.6%）。23.68%的参与者发展为重症COVID-19。总体而言，与covid -19相关的死亡率为15.26%。ABO血型间生存经验无差异。ABO血型与COVID-19严重程度和死亡率之间无统计学意义的关联。结论ABO血型差异与COVID-19严重程度和死亡率无相关性。此外，建议进行设计良好的对照研究，以探索ABO血型与COVID-19严重程度和死亡率的潜在联系。

{"title":"Do certain blood groups increase COVID-19 severity and mortality?","authors":"Tegene Atamenta kitaw, Ribka Nigatu Haile","doi":"10.1016/j.gloepi.2025.100212","DOIUrl":"10.1016/j.gloepi.2025.100212","url":null,"abstract":"<div><h3>Background</h3><div>The existence of a relationship between the ABO blood group and COVID-19 severity and mortality is still an unresolved concern. Some studies report that groups O and A show a lower and higher risk of developing severe COVID-19 and mortality, respectively. Some studies also report the reverse. There are inconclusive results from different studies. Thus, this study sought to determine the possible associations of ABO blood type with COVID-19 severity and mortality.</div></div><div><h3>Methods</h3><div>A retrospective study was conducted among 570 adults with real-time reverse transcription-polymerase chain reaction confirmed positive COVID-19 patients attending Eka Kotebe General Hospital, COVID-19 Treatment Center. A Kaplan-Meier survival curve was computed to examine the difference in survival experience between ABO blood groups. Multinomial and binary logistic regression models were fitted to determine the association between ABO blood group with COVID-19 severity and mortality, respectively.</div></div><div><h3>Result</h3><div>238 (41.8 %) COVID-19 patients had blood group B, followed by 201 (35.3 %) A, 82 (14.4 %) O, and 49 (8.6 %) AB blood type. 23.68 % of participants develop severe COVID-19. Overall, 15.26 % COVID-19-related mortality was found. No difference in survival experience was observed between ABO blood types. There was no statistically significant association between ABO blood type and COVID-19 severity, and mortality.</div></div><div><h3>Conclusion</h3><div>We found no relationship between ABO blood group differences and COVID-19 severity, and mortality. Further, well-design-controlled studies are suggested to explore the potential link of ABO blood group with COVID-19 severity and mortality.</div></div>","PeriodicalId":36311,"journal":{"name":"Global Epidemiology","volume":"10 ","pages":"Article 100212"},"PeriodicalIF":0.0,"publicationDate":"2025-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144605869","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Epidemiology of stroke in Pakistan and its provinces, 1990–2021: Findings from the global burden of disease study 2021 1990-2021年巴基斯坦及其各省中风流行病学：来自2021年全球疾病负担研究的结果

Global Epidemiology

Pub Date : 2025-07-05 DOI: 10.1016/j.gloepi.2025.100211

Sufyan Shahid , Ali Dheyaa Marsool , Maha Sajjad , Muneeb Saifullah , Muhammad Awais Alam , Syed Ijlal Ahmed , Raheel Ahmed , Fahd Sultan

Background

Despite stroke being a leading cause of death in Pakistan, no study has comprehensively analyzed national and regional trends over time. This gap limits understanding of the evolving burden and underlying risk factors.

Methods

We analyzed regional stroke trends in Pakistan from 1990 to 2021 using Global Burden of Disease (GBD) 2021 data. Age-standardized incidence, prevalence, mortality, disability-adjusted life years (DALYs), and 18 modifiable risk factors were analyzed across provinces. Joinpoint regression identified temporal trends through annual percent changes (APCs), and population attributable fractions (PAFs) quantified risk contributions. All estimates incorporated 95 % uncertainty intervals and underwent sensitivity testing to ensure robustness.

Findings

In 2021, stroke was the second leading cause of death in Pakistan, responsible for 99,759 deaths, with an incidence of 153 and a prevalence of 1088 per 100,000 population. Age-standardized rates for incidence, prevalence, mortality, and DALYs declined overall from 1990 to 2021 (ASIR AAPC: −0.15 %), though reductions were modest. Regional trends showed the steepest declines in Islamabad and Punjab, while Balochistan experienced an increase in mortality (ASMR AAPC: +0.03 %) and DALYs (+0.05 %). Sex-specific trends showed greater declines in incidence and mortality among females, while DALY reductions were modest in both sexes, reflecting ongoing disability. High systolic blood pressure was the leading risk factor for stroke mortality (25.2 per 100,000) and DALYs (674.9 per 100,000), followed by household air pollution, ambient particulate matter, and high fasting glucose. These findings reflect the combined influence of cardiometabolic, environmental, and behavioral risks on Pakistan's evolving stroke burden.

Interpretation

Despite modest improvements, stroke remains a major health challenge in Pakistan, with rising long-term disability and stark regional inequalities. Strengthening prevention, equitable access, and post-stroke care is essential to curb the growing burden.

尽管中风是巴基斯坦的主要死亡原因，但没有一项研究全面分析了国家和地区的长期趋势。这一差距限制了对不断变化的负担和潜在风险因素的理解。方法：我们使用全球疾病负担（GBD） 2021数据分析1990年至2021年巴基斯坦区域卒中趋势。对各省的年龄标准化发病率、患病率、死亡率、残疾调整生命年（DALYs）和18个可改变的危险因素进行分析。接合点回归通过年百分比变化（APCs）确定时间趋势，而人口归因分数（PAFs）量化风险贡献。所有的估计都包含95%的不确定区间，并经过敏感性测试以确保稳健性。2021年，中风是巴基斯坦第二大死因，造成99,759人死亡，发病率为153人，患病率为每10万人1088人。从1990年到2021年，年龄标准化的发病率、患病率、死亡率和DALYs总体下降（ASIR AAPC: - 0.15%），尽管下降幅度不大。区域趋势显示，伊斯兰堡和旁遮普的下降幅度最大，而俾路支省的死亡率和伤残调整生命年分别上升了0.03%和0.05%。按性别区分的趋势显示，女性的发病率和死亡率下降幅度较大，而男女伤残津贴年的下降幅度不大，反映出持续的残疾。高收缩压是中风死亡率（25.2 / 10万）和伤残调整生命年（674.9 / 10万）的主要危险因素，其次是家庭空气污染、环境颗粒物和空腹血糖过高。这些发现反映了心脏代谢、环境和行为风险对巴基斯坦不断变化的中风负担的综合影响。尽管有所改善，但中风仍然是巴基斯坦面临的主要健康挑战，长期残疾人数不断上升，地区不平等现象严重。加强预防、公平获取和卒中后护理对于遏制日益加重的负担至关重要。

{"title":"Epidemiology of stroke in Pakistan and its provinces, 1990–2021: Findings from the global burden of disease study 2021","authors":"Sufyan Shahid , Ali Dheyaa Marsool , Maha Sajjad , Muneeb Saifullah , Muhammad Awais Alam , Syed Ijlal Ahmed , Raheel Ahmed , Fahd Sultan","doi":"10.1016/j.gloepi.2025.100211","DOIUrl":"10.1016/j.gloepi.2025.100211","url":null,"abstract":"<div><h3>Background</h3><div>Despite stroke being a leading cause of death in Pakistan, no study has comprehensively analyzed national and regional trends over time. This gap limits understanding of the evolving burden and underlying risk factors.</div></div><div><h3>Methods</h3><div>We analyzed regional stroke trends in Pakistan from 1990 to 2021 using Global Burden of Disease (GBD) 2021 data. Age-standardized incidence, prevalence, mortality, disability-adjusted life years (DALYs), and 18 modifiable risk factors were analyzed across provinces. Joinpoint regression identified temporal trends through annual percent changes (APCs), and population attributable fractions (PAFs) quantified risk contributions. All estimates incorporated 95 % uncertainty intervals and underwent sensitivity testing to ensure robustness.</div></div><div><h3>Findings</h3><div>In 2021, stroke was the second leading cause of death in Pakistan, responsible for 99,759 deaths, with an incidence of 153 and a prevalence of 1088 per 100,000 population. Age-standardized rates for incidence, prevalence, mortality, and DALYs declined overall from 1990 to 2021 (ASIR AAPC: −0.15 %), though reductions were modest. Regional trends showed the steepest declines in Islamabad and Punjab, while Balochistan experienced an increase in mortality (ASMR AAPC: +0.03 %) and DALYs (+0.05 %). Sex-specific trends showed greater declines in incidence and mortality among females, while DALY reductions were modest in both sexes, reflecting ongoing disability. High systolic blood pressure was the leading risk factor for stroke mortality (25.2 per 100,000) and DALYs (674.9 per 100,000), followed by household air pollution, ambient particulate matter, and high fasting glucose. These findings reflect the combined influence of cardiometabolic, environmental, and behavioral risks on Pakistan's evolving stroke burden.</div></div><div><h3>Interpretation</h3><div>Despite modest improvements, stroke remains a major health challenge in Pakistan, with rising long-term disability and stark regional inequalities. Strengthening prevention, equitable access, and post-stroke care is essential to curb the growing burden.</div></div>","PeriodicalId":36311,"journal":{"name":"Global Epidemiology","volume":"10 ","pages":"Article 100211"},"PeriodicalIF":0.0,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144579735","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Explainable artificial intelligence for predicting dengue outbreaks in Bangladesh using eco-climatic triggers 利用生态气候触发器预测孟加拉国登革热疫情的可解释人工智能

Global Epidemiology

Pub Date : 2025-06-05 DOI: 10.1016/j.gloepi.2025.100210

Md. Siddikur Rahman, Md. Abu Bokkor Shiddik

Background

Dengue represents a significant public health threat in Bangladesh, characterized by its complex ecological transmission dynamics. To improve dengue prevention and control efforts, firstly, we employ state-of-the-art artificial intelligence (AI) methods to identify the roles of eco-climatic factors in predicting dengue outbreaks in Bangladesh.

Methods

We utilize high-performance machine learning (ML) models, XGBoost and LightGBM, combined with explainable AI (XAI) methodologies to evaluate the predictive performance and impact of various dengue determinants in Bangladesh from 2000 to 2023. The LightGBM and XGBoost models were also utilized to predict dengue cases and early warning trends from 2024 to 2030. Climatic, socio-demographic, and landscape features were used to train these models; SHapley Additive Explanations (SHAP) values and LIME (Local Interpretable Model-agnostic Explanations) were used to interpret the results.

Findings

Between 2000 and 2023, Bangladesh experienced the highest number of dengue cases in August, while November saw the most fatalities. The XGBoost model excelled in predicting dengue outbreaks, achieving an AUC score of 0.89, a Log Loss of 0.64. Key predictors identified by the model include population density, precipitation, temperature, and land-use patterns. Additionally, Local Interpretable Model-agnostic Explanations (LIME) provided insights into the model's predictions, highlighting the significance of population density, relative humidity, and minimum temperature in dengue outbreaks.

Interpretation

This study showcases the potential of XAI in uncovering the complexities of dengue outbreaks, providing a robust tool for public health interventions. Our AI-driven framework can be utilized to generate prompt and timely alerts to prevent imminent dengue and other infectious disease outbreaks.

登革热在孟加拉国是一个重大的公共卫生威胁，其特点是其复杂的生态传播动态。为了改善登革热预防和控制工作，首先，我们采用最先进的人工智能（AI）方法来确定生态气候因素在预测孟加拉国登革热疫情中的作用。方法利用高性能机器学习（ML）模型，XGBoost和LightGBM，结合可解释人工智能（XAI）方法，评估2000年至2023年孟加拉国各种登革热决定因素的预测性能和影响。利用LightGBM和XGBoost模型预测2024 - 2030年登革热病例和预警趋势。使用气候、社会人口和景观特征来训练这些模型；使用SHapley加性解释（SHAP）值和LIME （Local Interpretable Model-agnostic Explanations）来解释结果。在2000年至2023年期间，孟加拉国8月份的登革热病例最多，而11月份的死亡人数最多。XGBoost模型在预测登革热暴发方面表现出色，AUC得分为0.89，Log Loss为0.64。该模型确定的关键预测因子包括人口密度、降水、温度和土地利用模式。此外，局部可解释模型不可知解释（LIME）为模型预测提供了见解，强调了人口密度、相对湿度和最低温度在登革热暴发中的重要性。本研究展示了XAI在揭示登革热暴发复杂性方面的潜力，为公共卫生干预提供了强有力的工具。我们的人工智能驱动框架可用于产生迅速和及时的警报，以预防即将发生的登革热和其他传染病暴发。

{"title":"Explainable artificial intelligence for predicting dengue outbreaks in Bangladesh using eco-climatic triggers","authors":"Md. Siddikur Rahman, Md. Abu Bokkor Shiddik","doi":"10.1016/j.gloepi.2025.100210","DOIUrl":"10.1016/j.gloepi.2025.100210","url":null,"abstract":"<div><h3>Background</h3><div>Dengue represents a significant public health threat in Bangladesh, characterized by its complex ecological transmission dynamics. To improve dengue prevention and control efforts, firstly, we employ state-of-the-art artificial intelligence (AI) methods to identify the roles of eco-climatic factors in predicting dengue outbreaks in Bangladesh.</div></div><div><h3>Methods</h3><div>We utilize high-performance machine learning (ML) models, XGBoost and LightGBM, combined with explainable AI (XAI) methodologies to evaluate the predictive performance and impact of various dengue determinants in Bangladesh from 2000 to 2023. The LightGBM and XGBoost models were also utilized to predict dengue cases and early warning trends from 2024 to 2030. Climatic, socio-demographic, and landscape features were used to train these models; SHapley Additive Explanations (SHAP) values and LIME (Local Interpretable Model-agnostic Explanations) were used to interpret the results.</div></div><div><h3>Findings</h3><div>Between 2000 and 2023, Bangladesh experienced the highest number of dengue cases in August, while November saw the most fatalities. The XGBoost model excelled in predicting dengue outbreaks, achieving an AUC score of 0.89, a Log Loss of 0.64. Key predictors identified by the model include population density, precipitation, temperature, and land-use patterns. Additionally, Local Interpretable Model-agnostic Explanations (LIME) provided insights into the model's predictions, highlighting the significance of population density, relative humidity, and minimum temperature in dengue outbreaks.</div></div><div><h3>Interpretation</h3><div>This study showcases the potential of XAI in uncovering the complexities of dengue outbreaks, providing a robust tool for public health interventions. Our AI-driven framework can be utilized to generate prompt and timely alerts to prevent imminent dengue and other infectious disease outbreaks.</div></div>","PeriodicalId":36311,"journal":{"name":"Global Epidemiology","volume":"10 ","pages":"Article 100210"},"PeriodicalIF":0.0,"publicationDate":"2025-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144241899","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Artificial intelligence for advancing eye care in resource-poor settings: Assessing the predictive accuracy of an AI-model for diabetic retinopathy screening in India 在资源贫乏环境中推进眼科保健的人工智能：评估印度糖尿病视网膜病变筛查人工智能模型的预测准确性

Global Epidemiology

Pub Date : 2025-06-01 DOI: 10.1016/j.gloepi.2025.100209

Rohan Chawla , Prachi Karkhanis , Malay Shah , Aritra Das , Rishabh Sharma , Dhwani Almaula , Pradeep Venkatesh , Harsh Vardhan Singh , Mukul Kumar , Ramanuj Samanta , Vinod Kumarl , Amar Shah , Bhavin Vadera , Nakul Jain , Akanksha Sen , Shyamsundar Shreedhar , Vipin Garg , Soma Dhaval , Kowshik Ganesh , Srinivas Rana , Radhika Tandon

Background

Timely identification and treatment of Diabetic Retinopathy (DR) is critical in avoiding vision loss. DR screening is challenging, especially in resource-limited areas where trained ophthalmologists are scarce. AI solutions show promise in addressing this challenge. In this study, the performance metrics of an AI solution (MadhuNetrAI) developed in India was evaluated for referring and grading DR.

Methods

MadhuNetrAI was developed de novo by the All India Institute of Medical Sciences (AIIMS) and Wadhwani AI (WIAI). It was tested on 1078 fundus images (from AIIMS Delhi and an unannotated subset of publicly available EyePACS images) against two ophthalmologists and an adjudicator serving as independent gold-standard annotators, wherein the disease status of the patients remained unknown.

Findings

MadhuNetrAI demonstrated high sensitivity (93·2 %; CI: 89·5 %–95·6 %) and specificity (95·3 %; CI: 93·7 %–96·6 %) in detecting referable DR (moderate, severe, proliferative DR). The area-under-the-curve for referring DR against the gold standard was 0·97 (CI: 0·95–0·99) indicating excellent diagnostic performance. The agreement in grading DR severity was high (kappa = 0·89, CI: 0·86–0·91). The model performed comparably in detecting DR too.

Interpretation

MadhuNetrAI's ability to grade DR severity and identify referrable cases could bring DR patients to care much earlier. Further research and clinical trials are needed to ensure its reliability and generalizability across diverse populations and image qualities.

Funding

MadhuNetrAI was developed by technical and programmatic teams at WIAI, with inputs and contributions by the clinical team at AIIMS, and funded by USAID. The authors have no financial or non-financial conflicts of interest to disclose.

背景及时发现和治疗糖尿病视网膜病变（DR）是避免视力丧失的关键。DR筛查具有挑战性，特别是在资源有限、训练有素的眼科医生稀缺的地区。人工智能解决方案有望解决这一挑战。在本研究中，对印度开发的人工智能解决方案（MadhuNetrAI）的性能指标进行评估，以参考和评分博士。方法MadhuNetrAI由全印度医学科学研究所（AIIMS）和Wadhwani AI （WIAI）重新开发。在1078张眼底图像（来自AIIMS Delhi和公开可用的EyePACS图像的未注释子集）上，对两名眼科医生和一名作为独立金标准注释者的审稿人进行了测试，其中患者的疾病状态仍然未知。发现smadhunetrai具有高灵敏度(93.2%；CI: 89.5% - 95.6%)和特异性(95.3%；CI: 93.7% - 96%)在发现可参考DR（中度、重度、增殖性DR）方面。参照金标准的DR曲线下面积为0.97 (CI: 0.95 ~ 0.99)，诊断效果良好。对DR严重程度分级的一致性较高（kappa = 0.89, CI: 0.86 ~ 0.91）。该模型在检测DR方面也有相当的效果。madhunetrai分级DR严重程度和识别转诊病例的能力可以使DR患者更早接受治疗。需要进一步的研究和临床试验来确保其在不同人群和图像质量中的可靠性和普遍性。madhunetrai由WIAI的技术和项目团队开发，AIIMS的临床团队提供投入和贡献，并由美国国际开发署资助。作者没有财务或非财务利益冲突要披露。

{"title":"Artificial intelligence for advancing eye care in resource-poor settings: Assessing the predictive accuracy of an AI-model for diabetic retinopathy screening in India","authors":"Rohan Chawla , Prachi Karkhanis , Malay Shah , Aritra Das , Rishabh Sharma , Dhwani Almaula , Pradeep Venkatesh , Harsh Vardhan Singh , Mukul Kumar , Ramanuj Samanta , Vinod Kumarl , Amar Shah , Bhavin Vadera , Nakul Jain , Akanksha Sen , Shyamsundar Shreedhar , Vipin Garg , Soma Dhaval , Kowshik Ganesh , Srinivas Rana , Radhika Tandon","doi":"10.1016/j.gloepi.2025.100209","DOIUrl":"10.1016/j.gloepi.2025.100209","url":null,"abstract":"<div><h3>Background</h3><div>Timely identification and treatment of Diabetic Retinopathy (DR) is critical in avoiding vision loss. DR screening is challenging, especially in resource-limited areas where trained ophthalmologists are scarce. AI solutions show promise in addressing this challenge. In this study, the performance metrics of an AI solution (MadhuNetrAI) developed in India was evaluated for referring and grading DR.</div></div><div><h3>Methods</h3><div>MadhuNetrAI was developed de novo by the All India Institute of Medical Sciences (AIIMS) and Wadhwani AI (WIAI). It was tested on 1078 fundus images (from AIIMS Delhi and an unannotated subset of publicly available EyePACS images) against two ophthalmologists and an adjudicator serving as independent gold-standard annotators, wherein the disease status of the patients remained unknown.</div></div><div><h3>Findings</h3><div>MadhuNetrAI demonstrated high sensitivity (93·2 %; CI: 89·5 %–95·6 %) and specificity (95·3 %; CI: 93·7 %–96·6 %) in detecting referable DR (moderate, severe, proliferative DR). The area-under-the-curve for referring DR against the gold standard was 0·97 (CI: 0·95–0·99) indicating excellent diagnostic performance. The agreement in grading DR severity was high (kappa = 0·89, CI: 0·86–0·91). The model performed comparably in detecting DR too.</div></div><div><h3>Interpretation</h3><div>MadhuNetrAI's ability to grade DR severity and identify referrable cases could bring DR patients to care much earlier. Further research and clinical trials are needed to ensure its reliability and generalizability across diverse populations and image qualities.</div></div><div><h3>Funding</h3><div>MadhuNetrAI was developed by technical and programmatic teams at WIAI, with inputs and contributions by the clinical team at AIIMS, and funded by USAID. The authors have no financial or non-financial conflicts of interest to disclose.</div></div>","PeriodicalId":36311,"journal":{"name":"Global Epidemiology","volume":"9 ","pages":"Article 100209"},"PeriodicalIF":0.0,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144231531","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0