Frontiers in Public Health最新文献

{"title":"Effects of urban green exercise on mental health: a systematic review and meta-analysis.","authors":"Guidan Hu, Qingyuan Luo, Peng Zhang, Hao Zeng, Xiujie Ma","doi":"10.3389/fpubh.2025.1677223","DOIUrl":"10.3389/fpubh.2025.1677223","url":null,"abstract":"<p><p>As global urbanization accelerates, concerns regarding the mental health of urban residents have become increasingly prominent. Urban green exercise, a non-pharmacological intervention integrating exposure to nature with physical activity, has gained considerable attention due to its potential mental health benefits. However, systematic evidence synthesizing the specific effects and underlying mechanisms of urban green exercise on mental health remains limited. Following strict adherence to PRISMA guidelines, systematic searches of PubMed, MEDLINE, Embase, the Cochrane Library, and Web Of Science identified 15 RCTs involving urban green spaces, comprising 980 participants aged 18 years and older. Methodological quality was evaluated using the Cochrane RoB 2.0 tool. Meta-analysis using standardized mean difference (SMD) was conducted using Stata 17.0, while subgroup and regression analyses were performed to explore moderating factors, including intervention period, frequency, duration per session, exercise intensity (METs), and gender. A moderate and statistically significant positive impact of urban green exercise on mental health was found (SMD = -0.40; 95% CI = -0.56 to -0.25; <i>p</i> < 0.001), with low between-study heterogeneity (I² = 33.9%). The most pronounced effects were associated with short-duration interventions (<12 weeks), interventions conducted at least three times weekly, session durations of 20 min or less, and low-to-moderate intensity (≤3 METs). Greater benefits were observed among female participants. This systematic review provides robust empirical support for the mental health benefits of urban green exercise in adult urban populations, highlighting the efficacy of short-duration, high-frequency, moderate-intensity intervention models. These findings offer evidence-based insights to inform urban public health policy and green space planning, emphasizing the need to enhance the accessibility and utilization of urban green spaces. Further high-quality RCTs with larger sample sizes are recommended to further validate long-term effects and elucidate underlying mechanisms.</p>","PeriodicalId":12548,"journal":{"name":"Frontiers in Public Health","volume":"13 ","pages":"1677223"},"PeriodicalIF":3.4,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12507887/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145279624","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Global, regional, and national total burden related to hepatitis B and attributable risk factors in adults aged 65 years and older from 1990 to 2021 and projection to 2030.","authors":"Linying Gao, Yanfeng Ren, Songyue Hu, Yifan Zhang, Hongjing Bai, Jinbo Li, Keke Wang, Suping Wang, Yongliang Feng","doi":"10.3389/fpubh.2025.1654356","DOIUrl":"10.3389/fpubh.2025.1654356","url":null,"abstract":"<p><strong>Background: </strong>Hepatitis B virus and its complications remain a major global public health threat. This study aimed to assess the global, regional, and national trends of total burden related to hepatitis B among older adults aged ≥65 years from 1990 to 2021 and predict trend to 2030.</p><p><strong>Methods: </strong>Drawing upon the Global Burden of Diseases (GBD) 2021, we obtained the data on incidence, prevalence, mortality, and disability-adjusted life years (DALYs) from 1990 to 2021, categorized by sex, age, and socio-demographic index (SDI). Associations of. these metrics and SDI were analyzed. Secular trends were evaluated employing average annual percentage changes (AAPC) and joinpoint regression, with future trajectories estimated through 2030 using the Bayesian age-period-cohort (BAPC) model.</p><p><strong>Results: </strong>From 1990 to 2021, the global age-standardized incidence rate (ASIR) of total burden related to hepatitis B among older adults aged ≥65 years declined from 710.53 to 591.79 per 100,000 population (AAPC -0.57%), and age-standardized DALYs rate (ASDR) declined from 927.65 to 605.93 per 100,000 population (AAPC -1.39%). ASIR and ASPR both showed significant decreasing trends in males and females, with persistently higher rates observed among males across all study years. The ASIR increased only in the >95 years age group, while a significant increase in DALYs was observed across all age strata, with the magnitude of increase escalating with advancing age. ASIR increased exclusively in high SDI countries (AAPC 0.16%). In 2021, Central Sub-Saharan Africa accounted for the peak ASIR (1,671.36 per 100,000 population), conversely North Africa and the Middle East showed the most pronounced ASDR decline (AAPC -2.74%). Both alcohol use and smoking emerged as the most consequential modifiable risk factors for DALYs in adults aged ≥65 years. Projections indicate declining trends in ASIR, ASDR, and ASMR for this population by 2030, whereas ASPR is projected to exhibit an upward trend.</p><p><strong>Conclusion: </strong>The global burden and trends of total burden related to hepatitis B among older adults aged ≥65 years showed a significant decline, with projections indicating continued reduction by 2030. However, notable regional and national disparities persist. Key challenges in managing health risks for adults aged ≥65 years remain centered on alcohol use and smoking.</p>","PeriodicalId":12548,"journal":{"name":"Frontiers in Public Health","volume":"13 ","pages":"1654356"},"PeriodicalIF":3.4,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12507896/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145279696","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Interactive effects between lead-cadmium co-exposure and VEGFA gene polymorphisms on renal dysfunction: a gene-environment interaction study.","authors":"Yaotang Deng, Yunman Wen, Weixia Duan, Zhiqiang Zhao, Guoliang Li, Jiazhen Zhou, Le Yang, Jieyi Yang, Yapei Sun, Manyi Qiu, Lili Liu","doi":"10.3389/fpubh.2025.1627634","DOIUrl":"10.3389/fpubh.2025.1627634","url":null,"abstract":"<p><strong>Background: </strong>Lead (Pb) and cadmium (Cd) are common persistent environmental pollutants, may cause renal dysfunction following long-term exposure. This study investigated whether vascular endothelial growth factor a (VEGFA) gene polymorphisms modify the association between Pb and Cd exposure with renal dysfunction risk, given the key role of gene-environment interactions in kidney pathogenesis.</p><p><strong>Methods: </strong>A cross-sectional study of 408 workers was undertaken from a Pb-Cd smelter in Guangdong Province, China in 2023. Metals in blood and urine were measured using Inductively coupled plasma-Mass Spectrometry (ICP-MS). Additive, dominant and recessive genetic models were employed to analyze differences in genotype distribution of rs3025010, rs10434 and rs833061 between normal and renal dysfunction groups. Interaction analyses were conducted to examine the combined effects of blood lead (BPb) and urinary cadmium (UCd) exposure with these polymorphisms under different genetic models on renal dysfunction risk.</p><p><strong>Results: </strong>For rs833061 locus, BPb showed statistically significant differences in both the additive and recessive models (<i>p</i> < 0.05), while renal function exhibited differences in the additive and dominant models (<i>p</i> < 0.05). For rs3025010, BPb showed significant differences in the recessive model (<i>p</i> = 0.05), and renal function demonstrated differences in both additive and dominant models (<i>p</i> < 0.05). Multivariate regression analysis identified BPb and UCd as risk factors for renal dysfunction, with odds ratios ranging from 1.40 to 3.46 (<i>p</i> < 0.05). Interaction analyses revealed interactions between rs3025010 and Pb [BPb × rs3025010: OR (95%CI) = 0.69(0.49, 0.88)] in dominant model. The rs10434 locus interactions with Pb and Cd in both the additive [OR (95%CI) = 0.60 (0.31, 0.91)] and recessive models [OR (95%CI) = 0.51 (0.27, 0.85)].</p><p><strong>Conclusion: </strong>This study identified significant gene-environment interactions between VEGFA polymorphisms (rs3025010 and rs10434) and Pb-Cd co-exposure in renal dysfunction. These findings suggest that screening for these polymorphisms could identify high-risk populations for targeted prevention and control strategies.</p>","PeriodicalId":12548,"journal":{"name":"Frontiers in Public Health","volume":"13 ","pages":"1627634"},"PeriodicalIF":3.4,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12507740/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145279643","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Shared decision-making among older adults with multimorbidity in Kerala's primary care: a qualitative study using the socio-ecological model.","authors":"Abhijith A Kumar, Asha Kamath, Lena Ashok, Veena Ganesh Kamath, Varalakshmi Chandra Sekaran","doi":"10.3389/fpubh.2025.1665368","DOIUrl":"10.3389/fpubh.2025.1665368","url":null,"abstract":"<p><strong>Introduction: </strong>Multimorbidity is an increasing public health challenge among older adults, particularly in Kerala, India. Shared decision-making (SDM) is central to person-centered care for this group, yet its implementation remains inconsistent in low- and middle-income countries (LMICs).</p><p><strong>Methods: </strong>This qualitative study explored the experiences and perspectives of older adults with multimorbidity regarding SDM in Kerala's primary care. Sixteen adults (aged 60+) were recruited from four primary health centers using purposive sampling. The Socio-Ecological Model guided the design and thematic analysis, examining SDM influences at individual, interpersonal, organizational, and sociocultural levels.</p><p><strong>Results: </strong>Key findings revealed that individual barriers, such as limited health literacy and low self-efficacy, led to passive participation. Family members played a central role in healthcare interactions, sometimes facilitating but occasionally overshadowing patient voices. Organizational barriers, including high patient load and time constraints, limited SDM, while positive provider communication and continuity enabled engagement. Sociocultural factors included strong respect for medical authority and pluralistic health-seeking, with patients often reluctant to disclose alternative treatments to allopathic doctors. Exclusion from SDM was linked to dissatisfaction and poor adherence.</p><p><strong>Conclusion: </strong>Addressing these barriers and leveraging enablers will require coordinated efforts in communication, health literacy, family engagement, and culturally sensitive practice to advance person-centered care.</p>","PeriodicalId":12548,"journal":{"name":"Frontiers in Public Health","volume":"13 ","pages":"1665368"},"PeriodicalIF":3.4,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12509231/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145279727","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Assessment and prediction of copper release amount from copper oxide facepieces.","authors":"Zengqing Bai, Chenchen Sun, Jinyan Liu, Zenghui Liu","doi":"10.3389/fpubh.2025.1664838","DOIUrl":"10.3389/fpubh.2025.1664838","url":null,"abstract":"<p><strong>Background: </strong>Disposable facepieces, as important personal protective equipment, provide respiratory protection for workers. However, Cu containing facepieces may cause Cu release, posing a potential danger to human health.</p><p><strong>Methods: </strong>In this study, aging experiments were conducted on 36 groups of facepieces, simulating the use of facepieces under high temperature, radiation environment and work rate to assess the exposure levels of workers to Cu amount. Meanwhile, a machine learning model was developed based on the Cu release amount to predict the exposure level.</p><p><strong>Results: </strong>The research found that after simulating the aging of facepieces, the Cu release ranged from 7.25µg to 23.65µg, and the release trend showed an increasing trend under the simulated harsh conditions. The exposure levels in different scenarios were evaluated based on the release amount. Among them, 27 groups were evaluated as level III and 9 groups were evaluated as level II. Furthermore, the prediction results of Support Vector Machine (SVM), Backpropagation Neural Network (BPNN), and Random Forest (RF), test and training sets were evaluated using coefficient of determination (R²), root mean square error (RMSE) and mean absolute error (MAE). Among them, the SVM algorithm performed the best, further improving its predictive ability by using data augmentation methods and Particle Swarm Optimization (R² of 0.9045, RMSE of 0.0762, and MAE of 0.0525). The relative errors between the predicted values and the true values of all samples were mostly less than 5%.</p><p><strong>Conclusion: </strong>The research method in this study can effectively assess the Cu exposure level of workers and provide a scientific basis for occupational health monitoring.</p>","PeriodicalId":12548,"journal":{"name":"Frontiers in Public Health","volume":"13 ","pages":"1664838"},"PeriodicalIF":3.4,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12507794/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145279761","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A cross-sectional study on the mediating role of psychological capital in the relationship between workplace violence and professional commitment among nursing students.","authors":"Zhengrong Cai, Lili Jiao, JinJing Huang, Cui Peng, AiLing Chen, ZhouMin Shen, Yan Liu","doi":"10.3389/fpubh.2025.1629330","DOIUrl":"10.3389/fpubh.2025.1629330","url":null,"abstract":"<p><strong>Background: </strong>Studies have documented associations among workplace violence (WPV), psychological capital (PsyCap), and professional commitment (PC). But limited research has investigated how PsyCap mediates the relationship between WPV and PC, particularly among nursing interns.</p><p><strong>Objective: </strong>This study assessed the current status of WPV, PsyCap, and PC among Chinese nursing students, analyzed thier interrelationships and further determined whether PsyCap mediated the association between WPV and PC.</p><p><strong>Methods: </strong>A cross-sectional survey of 520 nursing interns used validated instruments: a demographic questionnaire, the WPV Scale, the PC Scale, and the PsyCap Questionnaire (Chinese version of PCQ). Pearson correlation analyzed the relationships among PsyCap, WPV, and PC, while hierarchical regression tested the mediation model.</p><p><strong>Results: </strong>The total scores for PC and PsyCap among nursing students were (77.04 ± 17.04) and (91.90 ± 16.16), respectively. Approximately 31.54% of the participants reported experiencing WPV. PsyCap was inversely associated with WPV (<i>r</i> = -0.619, <i>p</i> = 0.024) and positively associated with PC (<i>r</i> = 0.620, <i>p</i> < 0.001), in contrast WPV showed a negative correlation with PC (<i>r</i> = -0.807, <i>p</i> = 0.005). Mediation analysis revealed that PsyCap mediated the WPV-PC relationship, accounting for 47.5% of the total effect.</p><p><strong>Conclusion: </strong>Nursing students exhibited moderate levels of PsyCap and PC, with the prevalence of WPV being slightly lower than reported in comparable studies. And PsyCap fully mediated the relationship between WPV and PC. These findings revealed that nursing administrators and educators should enhance and develop nursing students' PsyCap to lower the adverse effects of WPV and promote higher levels of PC.</p>","PeriodicalId":12548,"journal":{"name":"Frontiers in Public Health","volume":"13 ","pages":"1629330"},"PeriodicalIF":3.4,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12507840/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145279765","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Correction: Addressing the overuse of hospital emergency departments in the Portuguese NHS: a new paradigm.","authors":"Francisco Goiana-da-Silva, Soraia Costa, Filipa Malcata, Juliana Sá, Rafael Vasconcelos, Miguel Cabral, Raisa Guedes, Inês Morais-Vilaça, Lara Pinheiro-Guedes, João Ferreira, Nélson Pereira, José Gaspar-Pais, Judite Neves, Joaquim Monteiro, Vera Pires, Miguel Paiva, Rui Guimarães, Hutan Ashrafian, Rita Moreira, Fátima Fonseca, Filomena Cardoso, Jaime Alves, Ara Darzi, Fernando Araújo","doi":"10.3389/fpubh.2025.1628755","DOIUrl":"10.3389/fpubh.2025.1628755","url":null,"abstract":"<p><p>[This corrects the article DOI: 10.3389/fpubh.2024.1444951.].</p>","PeriodicalId":12548,"journal":{"name":"Frontiers in Public Health","volume":"13 ","pages":"1628755"},"PeriodicalIF":3.4,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12510260/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145279916","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"From the perspective of multi-theory model, factors influencing physical activity among community-dwelling older adults with type 2 diabetes in China: a mixed-methods study.","authors":"Panpan Huai, Bo Zhang, Jingjing Sun, Rui Xu, Linghui Zhang, Xiao Qiao, Weili Sun, Hui Yang, Jinli Guo, Huancheng Su","doi":"10.3389/fpubh.2025.1634294","DOIUrl":"10.3389/fpubh.2025.1634294","url":null,"abstract":"<p><strong>Aim: </strong>To examine the factors influencing physical activity among community-dwelling older adults with type 2 diabetes in China, and to provide a strong theoretical framework and empirical support for creating more individualized and scientific strategies for improving physical activity.</p><p><strong>Methods: </strong>The text was analyzed by the innovative combination of traditional thematic analysis method and topic modeling (python machine learning) through the qualitative study, aiming to deeply explore the experiences and views of older adults with type 2 diabetes in the community regarding physical activity. Quantitative study adopted a cross-sectional survey to objectively and efficiently discover causal relationships among data. Finally, the results from the two different researches were compared to identify differences, similarities, and contradictions, enabling mutual verification and supplementation of the research findings and compensating for the limitations of a single research method, thereby obtaining the ultimate results.</p><p><strong>Results: </strong>We compared and integrated the qualitative and quantitative study results, and finally determined 11 topics, including 23 key factors, as the set of the study results of this study.</p><p><strong>Conclusion: </strong>From the perspective of Multi-theory model, this study explored the related factors influencing the physical activity of older adults with type 2 diabetes in the community through a mixed-methods study. By comparing and integrating the results of qualitative and quantitative studies, we finally identified 11 topics, including 23 key factors, as the results of this study, such as \"Focus on physical activity itself,\" \"The persistence of physical activity,\" and \"Traditional conception,\" which were not mentioned in previous systematic literature search. These new discoveries provide a empirical support for creating more individualized and scientific strategies for improving physical activity behaviors among older adults with type 2 diabetes in the community.</p>","PeriodicalId":12548,"journal":{"name":"Frontiers in Public Health","volume":"13 ","pages":"1634294"},"PeriodicalIF":3.4,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12507910/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145279646","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Analysis of blood utilisation efficiency driven by clinical management and hospital heterogeneity.","authors":"Feng Wang, Zhi-Guo Xu, Ke Lv, Jingxian Fei, Haiying Yang, Zequn Ou, Yun Song, Jingxian Song, Yizhu Chen, Le Wang","doi":"10.3389/fpubh.2025.1668449","DOIUrl":"10.3389/fpubh.2025.1668449","url":null,"abstract":"<p><strong>Background: </strong>Blood is a critical yet scarce medical resource, and improving the efficiency with which it is utilised remains a major global challenge. In 2019, China introduced Quality Control Indicators for Clinical Blood Use in an attempt to standardise management. However, significant discrepancies remain between the intended policy and its practical implementation, resulting in inefficiencies and safety concerns.</p><p><strong>Objective: </strong>This study aims to quantitatively evaluate the impact of administrative policies on clinical blood use, identify the main factors affecting the efficiency with which blood is utilised, and analyse how hospital level and type influence transfusion practices.</p><p><strong>Methods: </strong>A retrospective, multicentre study was conducted using data from 24 secondary and tertiary hospitals in Huzhou between 2020 and 2024. Key quality control indicators and a 25-point transfusion record scoring system were employed. Trends were analysed using ANOVA and chi-square tests, and hospital stratification was analysed using MANOVA. Predictors of blood use per discharged patient were identified using multiple linear regression and linear mixed-effects models.</p><p><strong>Results: </strong>Over 5 years, the number of transfusion technicians increased by 72%, transfusion record scores improved by 34.6%, and per capita blood use decreased by 46.9%. However, blood use in low-complexity surgeries increased by an abnormal 200%. Tertiary hospitals showed higher blood use but better documentation than secondary hospitals. Regression analysis revealed that technician density (<i>β</i> = -0.280) and transfusion record score (<i>β</i> = -0.202) were negatively associated with blood use, whereas surgical complexity was positively associated with it. Hospital grade and type also significantly influenced outcomes.</p><p><strong>Conclusion: </strong>Efficiency in blood utilisation is more strongly influenced by process standardisation and human resources than by hospital level or type alone. Rather than rigid indicators, policy incentives should emphasise precision transfusion strategies and dynamic quality management to align resource use with clinical need and support sustainable blood management systems.</p>","PeriodicalId":12548,"journal":{"name":"Frontiers in Public Health","volume":"13 ","pages":"1668449"},"PeriodicalIF":3.4,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12507902/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145279806","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Artificial intelligence-driven prediction and interpretation of central line-associated bloodstream infections in ICU: insights from the MIMIC-IV database.","authors":"Yang He, Jiali Huang, Na Li, Gaosheng Zhou, Jinglan Liu","doi":"10.3389/fpubh.2025.1675077","DOIUrl":"10.3389/fpubh.2025.1675077","url":null,"abstract":"<p><strong>Objective: </strong>To develop and internally validate interpretable machine learning (ML) models for predicting individual central line-associated bloodstream infection (CLABSI) risk in adult ICU patients with central venous catheters (CVCs) using the MIMIC-IV database.</p><p><strong>Methods: </strong>We conducted a retrospective observational cohort study using the MIMIC-IV database. Adult ICU patients with both central venous catheter placement and blood culture evaluation were included. Patients were classified into CLABSI and non-CLABSI cohorts based on central venous catheter tip culture results. A comprehensive set of demographic, physiological, laboratory, therapeutic, and nursing variables was extracted. Feature selection employed Least Absolute Shrinkage and Selection Operator (LASSO) regression. Seven machine learning (ML) models-logistic regression, decision tree, random forest, XGBoost, support vector machine, neural network, and gradient boosting-were developed and compared. Discrimination and calibration were assessed using the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, specificity, F1 score, and Brier score. The optimal model was interpreted with SHAP (SHapley Additive exPlanations) values to elucidate feature contributions.</p><p><strong>Results: </strong>Among 11,999 ICU patients, 519 (4.3%) developed CLABSI. CLABSI patients were younger (61.0 vs. 66.0 years), had higher rates of multi-lumen catheters (91.3 vs. 63.6%), mechanical ventilation (90.9 vs. 74.0%), and dialysis (34.9 vs. 7.2%; all <i>p</i> < 0.001). The random forest model achieved optimal performance (AUC 0.950, 95% CI 0.931-0.966; sensitivity 0.904, specificity 0.865), outperforming traditional models. SHAP analysis identified ICU length of stay, unique caregivers, and arterial catheterization as top predictors. CLABSI cases exhibited prolonged ICU stays, increased caregiver exposure, and elevated inflammatory markers. Decision curve analysis confirmed clinical utility, with robust performance maintained in sensitivity analyses.</p><p><strong>Conclusion: </strong>Machine learning models, particularly the random forest model, accurately predict CLABSI risk in ICU patients. The use of interpretable AI techniques such as SHAP enhances transparency and provides actionable insights for clinical practice. These findings support the development of early warning systems to reduce CLABSI incidence and improve patient outcomes.</p>","PeriodicalId":12548,"journal":{"name":"Frontiers in Public Health","volume":"13 ","pages":"1675077"},"PeriodicalIF":3.4,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12507818/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145279804","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}