Health data science最新文献

{"title":"Association of Accelerometer-Measured Light-Intensity Physical Activity with Cardiovascular Diseases: A Large Cohort Study.","authors":"Lihui Zhou, Zhi Cao, Yujia Lv, Hongxi Yang, Dun Li, Jing Lin, Chenjie Xu, Yaogang Wang","doi":"10.34133/hds.0470","DOIUrl":"10.34133/hds.0470","url":null,"abstract":"<p><p><b>Background:</b> Light-intensity physical activity (LPA) constitutes a substantial portion of daily activities; however, it is more challenging to record and measure accurately compared to moderate- to vigorous-intensity physical activity (MVPA). Consequently, the health benefits associated with LPA are frequently overlooked. Our study aimed to investigate the associations of LPA pattern including objectively measured LPA's duration, proportion, and timing with risk of incident cardiovascular diseases (CVDs). <b>Methods:</b> With the UK Biobank, 83,658 participants free of CVD with valid accelerometry data were included between 2013 and 2015. Data on physical activity (PA) were collected using the Axivity AX3 wrist-based triaxial accelerometer worn for 1 week. LPA duration was estimated using machine learning models based on bouts data from wrist-worn accelerometers. The LPA proportion was calculated by dividing the amount of LPA by the sum of LPA and MVPA. LPA timing was determined using hierarchical clustering analysis based on the average duration of LPA within each time window across the 24-h cycle. CVDs of interest including coronary heart disease (CHD), stroke, heart failure (HF), and atrial fibrillation (AF) were identified through the linkage to hospital admission records and death registers. <b>Results:</b> Over a median follow-up duration of 7.98 (interquartile range: 7.43 to 8.47) years, a total of 2,239 CVDs, including 1,014 CHDs, 352 strokes, 796 AFs, and 194 HFs, were recorded. An L-shaped relationship was found between LPA duration and the risk of HF. Compared to individuals in the lowest quartile, those in the second (hazard ratio [HR], 0.64; 95% confidence interval [CI]: 0.44, 0.93), third (HR, 0.56; 95% CI: 0.37, 0.84), and highest quartile (HR, 0.65; 95% CI: 0.43, 0.99) of daily LPA were associated with lower risks of HF. Higher LPA in the morning, midday, and evening was associated with a higher risk of HF, whereas higher LPA in the night showed an adverse impact on HF risk. These findings were not observed for CHD, stroke, and AF. <b>Conclusions:</b> Our findings highlighted that maintaining sufficient duration of LPA might be beneficial for HF as a compensatory PA strategy, but it is also essential to regulate the proportion and timing of LPA.</p>","PeriodicalId":73207,"journal":{"name":"Health data science","volume":"6 ","pages":"0470"},"PeriodicalIF":4.8,"publicationDate":"2026-08-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13462243/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148726725","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Global Inequality and Injustice in Mortality Burden from Floods and Landscape Fire-Sourced Air Pollution.","authors":"Bingjie Zhou, Rongbin Xu, Zhengyu Yang, Yuming Guo","doi":"10.34133/hds.0417","DOIUrl":"10.34133/hds.0417","url":null,"abstract":"<p><p><b>Background:</b> Climate change amplifies the frequency and severity of landscape fire and floods, creating disproportionate health risks across countries. Yet, global disparities and injustices in mortality attributable to these climate-related risks remain poorly quantified. <b>Methods:</b> We constructed a country-level panel dataset (2000 to 2019) integrating mortality attributable to landscape fire-sourced (LFS) air pollution and floods, national Human Development Index (HDI), and per-capita CO₂ emissions. Relative inequality was assessed using the Gini and concentration indices, absolute inequality using HDI-based slope estimates, and climate injustice using the slope of attributable mortality rates against per-capita CO₂ emissions. <b>Results:</b> Mortality burdens from both risks were highly unequal across countries. LFS air pollution showed lower relative inequality (Gini = 0.59 for LFS air pollution and 0.82 for floods) but a steeper socioeconomic gradient (concentration index = -0.41 vs -0.28) and a larger absolute slope with HDI. Inequality from LFS air pollution declined modestly over time, whereas flood-related disparities remained less pronounced. Mortality was inversely associated with per-capita CO₂ emissions, indicating stronger and more persistent climate injustice for LFS air pollution than for floods. <b>Conclusion:</b> This study bridges data-driven analysis and policy perspectives to quantify global inequality and injustice in climate-related mortality burdens due to LFS air pollution and floods. This provides essential data to align emission mitigation and adaptive health policies toward greater environmental equity and justice under climate change.</p>","PeriodicalId":73207,"journal":{"name":"Health data science","volume":"6 ","pages":"0417"},"PeriodicalIF":4.8,"publicationDate":"2026-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13331477/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148391795","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Global Disease Burden Attributable to Ambient Air Pollution: Disparities, Determinants, and Implications for Public Health.","authors":"Yixuan Jiang, Su Shi, Xia Meng, Haidong Kan","doi":"10.34133/hds.0447","DOIUrl":"10.34133/hds.0447","url":null,"abstract":"<p><p><b>Background:</b> Ambient air pollution is a leading environmental risk factor for premature mortality worldwide. Characterizing the associated disease burden and spatiotemporal patterns is essential for informing global health policy. <b>Methods:</b> Using data from the Global Burden of Disease Study 2021, we assessed temporal trends and spatial heterogeneity in disease burden attributable to ambient fine particulate matter (PM_2.5) and ozone (O₃) during 1990-2020. We further examined the association between sociodemographic index (SDI) and disease burden, and conducted decomposition analysis to quantify the driving factors of temporal changes in disease burden. <b>Results:</b> From 1990 to 2020, the age-standardized death rates attributable to ambient PM_2.5 and O₃ both declined globally, but substantial regional heterogeneity was observed. High-SDI regions generally experienced declining burdens, whereas low- and middle-SDI regions showed persistent or increasing trends. A significant reversed U-shaped association between SDI and both PM_2.5- and O₃-attributable burden was observed, consistent with an environmental Kuznets-type pattern. Despite the decline in age-standardized rates, the absolute number of deaths attributable to ambient PM_2.5 and O₃ both increased, driven primarily by population aging, population growth, and increasing exposure levels, although reductions in baseline mortality rates partly offset these increases. <b>Conclusions:</b> Ambient air pollution continues to contribute to substantial and persistent disparities in global disease burden. Health-oriented and region-specific air pollution control strategies are needed to effectively mitigate these inequities and improve population health.</p>","PeriodicalId":73207,"journal":{"name":"Health data science","volume":"6 ","pages":"0447"},"PeriodicalIF":4.8,"publicationDate":"2026-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13331478/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148392699","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Beyond the Individual: A Data-Driven Approach to Protect Kidneys from Environmental Change.","authors":"Feifei Zhang, Luxia Zhang","doi":"10.34133/hds.0405","DOIUrl":"10.34133/hds.0405","url":null,"abstract":"","PeriodicalId":73207,"journal":{"name":"Health data science","volume":"6 ","pages":"0405"},"PeriodicalIF":4.8,"publicationDate":"2026-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13331476/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148392718","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Aligning Climate Mitigation, Air Pollution Control, and Healthy Aging.","authors":"Huali Wang, Tao Xue","doi":"10.34133/hds.0408","DOIUrl":"10.34133/hds.0408","url":null,"abstract":"","PeriodicalId":73207,"journal":{"name":"Health data science","volume":"6 ","pages":"0408"},"PeriodicalIF":4.8,"publicationDate":"2026-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13331475/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148392683","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrating Electronic Health Records and Large Language Models for Coarse-to-Fine Hybrid Disease Prediction.","authors":"Chutong Wang, Mingda Li, Xuebing Yang, Mengxuan Sun, Wensheng Zhang","doi":"10.34133/hds.0466","DOIUrl":"10.34133/hds.0466","url":null,"abstract":"<p><p><b>Background:</b> Electronic health records (EHRs) have been widely adopted in recent years. However, diagnosis codes in EHRs are often high-dimensional, abstract, and follow a long-tailed distribution, limiting the effectiveness and interpretability of traditional data-driven models for disease prediction. Large language models (LLMs) offer a promising alternative due to their broad knowledge and strong reasoning capabilities, yet their direct application remains challenging due to the vast prediction space and limited domain adaptation. <b>Methods:</b> To address these challenges, we propose a coarse-to-fine hybrid disease prediction framework. First, an EHR-driven base prediction module is used to generate coarse predictions. Second, patient information is serialized using a tailored prompt template, and a fine-tuned LLM further assesses the presence of high-probability candidate diseases. Finally, an aggregation module integrates outputs from both modules to yield refined predictions. <b>Results:</b> The coarse-to-fine hybrid disease prediction framework consistently outperforms existing non-LLM and LLM-based approaches in disease prediction. Specifically, compared to the second-best baselines, improvements of 1.6% on P@5, 1.1% on P@10, 0.7% on P@20, 0.3% on R@5, 1.6% on R@10, 1.0% on R@20, 1.4% on w-F1, and 1.0% on Jaccard for the eICU Collaborative Research Database and improvements of 2.6% on P@5, 4.3% on P@10, 4.2% on P@20, 4.2% on R@5, 5.9% on R@10, 4.5% on R@20, 6.0% on w-F1, and 2.5% on Jaccard for the Medical Information Mart for Intensive Care IV are achieved, respectively. Further analyses demonstrate the effectiveness of each component of the framework. <b>Conclusion:</b> By integrating EHR-driven base prediction with LLM-based assessments in a coarse-to-fine manner, this study provides an effective and practical disease-prediction framework for clinical decision support.</p>","PeriodicalId":73207,"journal":{"name":"Health data science","volume":"6 ","pages":"0466"},"PeriodicalIF":4.8,"publicationDate":"2026-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13243798/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148213541","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Electronic Health Record-Based Machine Learning Model for Predicting Disease Activity in Patients with Rheumatoid Arthritis.","authors":"Xiaoying Zhang, Chun Li, Zelin Yun, Yi Zhao, Shengguang Li, Wenqiang Fan, Limin Ma, Xiangheng Meng, Ru Li, Fangmin Xu, Jing Yang, Zhanguo Li","doi":"10.34133/hds.0461","DOIUrl":"10.34133/hds.0461","url":null,"abstract":"<p><p><b>Background:</b> The use of initial clinical assessments to predict therapeutic outcomes via machine learning (ML) is a promising frontier in precision medicine. The study aims to construct ML models capable of predicting disease activity in patients with rheumatoid arthritis (RA), thereby optimizing clinical decision-making and treatment selection. <b>Methods:</b> This multicenter retrospective study analyzed electronic health records (EHRs) from 1,864 patients with RA across 5 tertiary hospitals in China between 2017 and 2022. The dataset from Peking University People's Hospital (PKUPH) was employed as the training and internal validation cohort, whereas data from 4 other centers were used for external validation. Longitudinal variables, including demographics, laboratory indices, and medication regimens, at baseline, 3-month, and 6-month follow-up were integrated to capture dynamic disease patterns. Four ML models were trained to predict disease status 6 months post-treatment, with the primary outcome defined as clinical remission (disease activity score in 28 joints with erythrocyte sedimentation rate ≤ 2.6). <b>Results:</b> The final analysis included 1,629 patients from PKUPH and 235 from 4 other tertiary hospitals. In the internal validation phase, the optimal model achieved an accuracy of 95.3% and an area under the receiver operating characteristic curve (AUROC) of 0.971, with sensitivity, specificity, positive predictive, and negative predictive values of 98.1%, 84.2%, 96.1%, and 91.8%, respectively. The model exhibited generalizability in external validation, presenting an accuracy of 87.3% and an AUROC of 0.922. Furthermore, in the multiclass task of stratifying patients into remission, low, moderate, or high disease activity, the deep neural network model showed an accuracy of 68.6% and AUROC of 0.860. <b>Conclusions:</b> Longitudinal clinical data extracted from EHRs can be effectively leveraged to develop prognostic models. This study confirms that deep learning approaches trained on large-scale multicenter cohorts can accurately predict disease trajectories in RA, offering a valuable tool for personalized patient management.</p>","PeriodicalId":73207,"journal":{"name":"Health data science","volume":"6 ","pages":"0461"},"PeriodicalIF":4.8,"publicationDate":"2026-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13243797/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148213512","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Digital and AI-Empowered Health Elements: An Integrated Pathway to Advancing Health.","authors":"Pengfei Li, Ying Zhou, Yan Li, Erdan Dong, Luxia Zhang","doi":"10.34133/hds.0468","DOIUrl":"10.34133/hds.0468","url":null,"abstract":"<p><p>Health outcomes arise from complex and evolving interactions among biological, behavioral, social, environmental, and technological influences. The social determinants of health framework have played a central role in establishing the importance of nonmedical drivers of health and continues to inform population health research and policy. As health data ecosystems expand and digital technologies become increasingly embedded in daily life, there is growing interest in approaches that can more explicitly represent dynamic, multidomain interactions shaping health in contemporary societies. In this Analysis, we introduce the concept of Health Elements, a framework that conceptualizes health as an emergent property distinct from complex additive effects arising from the interplay of biological, behavioral, social, environmental, and technological domains. We discuss how recent advances in data science and artificial intelligence create new opportunities to study these interactions, while highlighting the need for conceptual clarity, methodological rigor, and strong governance. We describe how Health Elements differ from and complement existing determinant-based models, outline principles for empirical operationalization, and consider implications for public health, clinical care, and policy. We also examine ethical, equity, and governance considerations relevant to integrated, artificial intelligence-enabled health systems. Finally, we identify key scientific questions to guide future research on when and how Health Elements approaches can advance health understanding and decision-making in an increasingly digital world.</p>","PeriodicalId":73207,"journal":{"name":"Health data science","volume":"6 ","pages":"0468"},"PeriodicalIF":4.8,"publicationDate":"2026-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13178507/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147965787","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Health in a Digital Age: From Determinants to Dynamic Systems.","authors":"Michelle A Williams","doi":"10.34133/hds.0465","DOIUrl":"10.34133/hds.0465","url":null,"abstract":"","PeriodicalId":73207,"journal":{"name":"Health data science","volume":"6 ","pages":"0465"},"PeriodicalIF":4.8,"publicationDate":"2026-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13178506/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147965841","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Review of Deep Learning Methods for Irregularly Sampled Medical Time Series Data.","authors":"Chenxi Sun, Moxian Song, Derun Cai, Baofeng Zhang, Hongyan Li, Shenda Hong","doi":"10.34133/hds.0456","DOIUrl":"10.34133/hds.0456","url":null,"abstract":"<p><p><b>Importance:</b> Medical time series constitute the largest data type in electronic health records and are often irregularly sampled in real-world clinical settings. Such irregularly sampled medical time series exhibit uneven time intervals, missing observations, and heterogeneous sampling rates, posing substantial challenges for deep learning models. <b>Highlights:</b> In this paper, from an irregularity-aware and data-centric perspective, we categorize existing deep learning methods for irregularly sampled medical time series into missing-data-based and raw-data-based approaches. We analyze their theoretical foundations and practical implications and conduct experiments on benchmark and real-world medical datasets to compare their strengths and limitations. <b>Conclusion:</b> Based on these analyses, we provide practical recommendations and discuss open problems and future research directions for modeling irregularly sampled medical time series.</p>","PeriodicalId":73207,"journal":{"name":"Health data science","volume":"6 ","pages":"0456"},"PeriodicalIF":4.8,"publicationDate":"2026-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13136615/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147846907","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}