JMIR Medical Informatics最新文献

{"title":"Investigating the Role of Wearable Devices in Facilitating Telehealth Adoption Among the Aging Population: Mediation Analysis of US National Data.","authors":"Ruijing Wang, Onur Asan, Ting Liao","doi":"10.2196/68559","DOIUrl":"10.2196/68559","url":null,"abstract":"<p><strong>Background: </strong>Telehealth adoption has grown significantly, presenting valuable opportunities for the aging population to access health care remotely. Despite evidence of its benefits in managing chronic conditions and promoting independence, many older adults remain hesitant to adopt telehealth, preferring traditional in-person visits even post pandemic. Current literature largely focuses on younger or general populations, overlooking the unique barriers faced by older adults, such as technology literacy and access disparities.</p><p><strong>Objective: </strong>This study investigates how telehealth adoption among the aging population is influenced and mediated by relevant factors, including the use of wearable devices, demographic factors, health conditions, and physical activity levels.</p><p><strong>Methods: </strong>A secondary analysis was conducted on the Health Information National Trends Survey (HINTS 6) data collected from March to November 2022. Of the 6252 respondents, 1596 older adults (≥65 years) were included. Telehealth adoption was defined using 2 survey items on receiving or being offered telehealth services. We construct regression and mediation analyses to understand the relationships between telehealth adoption and influential factors, including demographics, physical activity levels, health conditions, and the use of wearable devices.</p><p><strong>Results: </strong>We found that wearable device use, while not directly significant, plays a critical role in adoption when mediated by factors such as education, income, and general health. Specifically, higher levels of education and income increased the likelihood of telehealth adoption (P<.001), underscoring the importance of socioeconomic status. Additionally, rural versus urban residency emerged as a critical factor (P=.003), with rural residents demonstrating lower adoption rates, highlighting the accessibility and technology literacy barriers in these areas. Health conditions were inversely associated with telehealth adoption, suggesting that healthier individuals may perceive less need for telehealth services. The total effect of wearable use on telehealth adoption was significant (P=.007), with indirect effects via education (P<.001), income (P=.007), and health conditions (P=.004). The findings underscore the role of socioeconomic factors in influencing the adoption of health technologies.</p><p><strong>Conclusions: </strong>While wearable device use is associated with increased telehealth adoption among older adults, its effect operates primarily through mediating factors such as education, income, and health status. These findings suggest that addressing disparities in socioeconomic status and health literacy is critical to increasing telehealth engagement in aging populations.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e68559"},"PeriodicalIF":3.8,"publicationDate":"2025-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12331362/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144801055","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":"Machine Learning-Based Analysis of Lifestyle Risk Factors for Atherosclerotic Cardiovascular Disease: Retrospective Case-Control Study.","authors":"Hye-Jin Kim, Heeji Choi, Hyo-Jung Ahn, Seung-Ho Shin, Chulho Kim, Sang-Hwa Lee, Jong-Hee Sohn, Jae-Jun Lee","doi":"10.2196/74415","DOIUrl":"10.2196/74415","url":null,"abstract":"<p><strong>Background: </strong>The risk of developing atherosclerotic cardiovascular disease (ASCVD) varies among individuals and is related to a variety of lifestyle factors in addition to the presence of chronic diseases.</p><p><strong>Objective: </strong>We aimed to assess the predictive accuracy of machine learning (ML) models incorporating lifestyle risk behaviors for ASCVD risk using the Korean nationwide database.</p><p><strong>Methods: </strong>Using data from the Korea National Health and Nutrition Examination Survey, 5 ML algorithms were used for the prediction of high ASCVD risk: logistic regression (LR), support vector machine, random forest, extreme gradient boosting, and light gradient boosting models. ASCVD risk was assessed using the pooled cohort equations, with a high-risk threshold of ≥7.5% over 10 years. Among the 8573 participants aged 40-79 years, propensity score matching (PSM) was used to adjust for demographic confounders. We divided the dataset into a training and a test dataset in an 8:2 ratio. We also used bootstrapping to train the ML model with the area under the receiver operating characteristics curve score. Shapley additive explanations were used to identify the models' important variables in assessing high ASCVD risks. In sensitivity analysis, we additionally performed binary LR analysis, in which the ML model's results were consistent with the conventional statistical model.</p><p><strong>Results: </strong>Of the 8573 participants, 41.7% (n=3578) had high ASCVD risk. Before PSM, age and sex differed significantly between groups. PSM (1:1) yielded 1976 patients with balanced demographics. After PSM, the high ASCVD risk group had higher alcohol or tobacco use, lower omega-3 intake, higher BMI, less physical activity, and spent less time sitting. In 5 ML models, the extreme gradient boosting model showed the highest area under the receiver operating characteristics curve, indicating superior overall discrimination between high and low ASCVD risk groups. However, the light gradient boosting model demonstrated better performance in accuracy, recall, and F1-score. Variable importance analysis using Shapley additive explanations identified smoking and age as the strongest predictors, while BMI, sodium or omega-3 intake, and low-density lipoprotein cholesterol also had significant variables. Sensitivity analysis using multivariable LR analysis also confirmed these findings, showing that smoking, BMI, and low-density lipoprotein cholesterol increased ASCVD risk, whereas omega-3 intake and physical activity were associated with lower risk.</p><p><strong>Conclusions: </strong>Analyzing lifestyle behavioral factors in ASCVD risk with an ML model improves the predictive performance compared to traditional models. Personalized prevention strategies tailored to an individual's lifestyle can effectively reduce ASCVD risk.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e74415"},"PeriodicalIF":3.8,"publicationDate":"2025-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12330983/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144801056","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 Machine Learning-Based Prognostication Model Enhances Prediction of Early Hepatic Encephalopathy in Patients With Noncancer-Related Cirrhosis: Multicenter Longitudinal Cohort Study in Taiwan.","authors":"Hsin-Yu Chen, Yiu-Hua Cheng, Wei-Chung Yeh, Yi-Chuan Chen, Yi-Wen Tsai","doi":"10.2196/71229","DOIUrl":"10.2196/71229","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Hepatic encephalopathy (HE) contributes significantly to mortality among patients with liver cirrhosis. Early prediction of HE is essential for clinical decision-making, yet remains challenging-particularly in noncancer-related cirrhosis due to the unpredictable disease course.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Objective: &lt;/strong&gt;This study aimed to develop a novel machine learning (ML) model to improve early prediction of HE in patients with noncancer-related cirrhosis.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;A multicenter, retrospective cohort study was conducted from January 2010 to December 2017 across all Chang Gung Memorial Hospital branches in northern, middle, and southern Taiwan. We applied several ML models to evaluate HE predictability and compared their performance in the training dataset and testing dataset. Optimal sensitivity and specificity were determined using the Youden index. The best ML model was interpreted by the Shapley Additive Explanations plot.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;A total of 5878 patients with cirrhosis were included in the analysis, of whom 1187 (20.2%) subsequently developed HE. Compared to the non-HE group, patients with HE were older (median age 55, IQR 46-65 vs median age 54, IQR 44-66 years; P=.04) and had higher rates of hepatitis B virus infection (351/1187, 30% vs 961/4691, 20.5%; P&lt;.001), alcohol use (540/1187, 45.5% vs 1512/4691, 32.2%; P&lt;.001), sepsis (393/1187, 33.1% vs 792/4691, 16.9%; P&lt;.001), and mortality (425/1187, 35.8% vs 502/4691, 10.7%; P&lt;.001), along with distinct laboratory abnormalities reflecting liver dysfunction. Among the ML algorithms evaluated, the extreme gradient boosting algorithm demonstrated the highest predictive accuracy, achieving an area under the curve (AUC) of 0.86 (95% CI 0.83-0.88) in the testing dataset. This performance was significantly superior to that of the neural network (AUC 0.79, 95% CI 0.76-0.81; P&lt;.001), support vector machine (AUC 0.77, 95% CI 0.73-0.80; P&lt;.001), and the model for end-stage liver disease score (AUC 0.74, 95% CI 0.71-0.77; P&lt;.001). Using a probability threshold of 0.25, the extreme gradient boosting model demonstrated a sensitivity of 72% (95% CI 0.67-0.77), specificity of 80% (95% CI 0.78-0.82), a positive predictive value of 48% (95% CI 43-53), and a negative predictive value of 92% (95% CI 90-94) in the testing set. Comparable performance was observed in the training dataset, with a sensitivity of 80% (95% CI 0.77-0.83), specificity of 81% (95% CI 0.80-0.82), and a negative predictive value of 94% at the same threshold. The most influential predictive variables identified by the model included serum ammonia, aspartate transaminase, alanine transaminase, prothrombin time, and serum potassium.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Conclusions: &lt;/strong&gt;We developed a novel ML model for predicting HE in patients with noncancer-related cirrhosis. This model provides a practical guide to help physicians and these patients in shared decision-making regarding ","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e71229"},"PeriodicalIF":3.8,"publicationDate":"2025-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12327908/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144796264","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":"Predictive Modeling of Osteonecrosis of the Femoral Head Progression Using MobileNetV3_Large and Long Short-Term Memory Network: Novel Approach.","authors":"Gang Kong, Qi Zhang, Dan Liu, Jingbo Pan, Kegui Liu","doi":"10.2196/66727","DOIUrl":"10.2196/66727","url":null,"abstract":"<p><strong>Background: </strong>The assessment of osteonecrosis of the femoral head (ONFH) often presents challenges in accuracy and efficiency. Traditional methods rely on imaging studies and clinical judgment, prompting the need for advanced approaches. This study aims to use deep learning algorithms to enhance disease assessment and prediction in ONFH, optimizing treatment strategies.</p><p><strong>Objective: </strong>The primary objective of this research is to analyze pathological images of ONFH using advanced deep learning algorithms to evaluate treatment response, vascular reconstruction, and disease progression. By identifying the most effective algorithm, this study seeks to equip clinicians with precise tools for disease assessment and prediction.</p><p><strong>Methods: </strong>Magnetic resonance imaging (MRI) data from 30 patients diagnosed with ONFH were collected, totaling 1200 slices, which included 675 slices with lesions and 225 normal slices. The dataset was divided into training (630 slices), validation (135 slices), and test (135 slices) sets. A total of 10 deep learning algorithms were tested for training and optimization, and MobileNetV3_Large was identified as the optimal model for subsequent analyses. This model was applied for quantifying vascular reconstruction, evaluating treatment responses, and assessing lesion progression. In addition, a long short-term memory (LSTM) model was integrated for the dynamic prediction of time-series data.</p><p><strong>Results: </strong>The MobileNetV3_Large model demonstrated an accuracy of 96.5% (95% CI 95.1%-97.8%) and a recall of 94.8% (95% CI 93.2%-96.4%) in ONFH diagnosis, significantly outperforming DenseNet201 (87.3%; P<.05). Quantitative evaluation of treatment responses showed that vascularized bone grafting resulted in an average increase of 12.4 mm in vascular length (95% CI 11.2-13.6 mm; P<.01) and an increase of 2.7 in branch count (95% CI 2.3-3.1; P<.01) among the 30 patients. The model achieved an AUC of 0.92 (95% CI 0.90-0.94) for predicting lesion progression, outperforming traditional methods like ResNet50 (AUC=0.85; P<.01). Predictions were consistent with clinical observations in 92.5% of cases (24/26).</p><p><strong>Conclusions: </strong>The application of deep learning algorithms in examining treatment response, vascular reconstruction, and disease progression in ONFH presents notable advantages. This study offers clinicians a precise tool for disease assessment and highlights the significance of using advanced technological solutions in health care practice.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e66727"},"PeriodicalIF":3.8,"publicationDate":"2025-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12327698/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144796265","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":"Probability-Based Early Warning for Seasonal Influenza in China: Model Development Study.","authors":"Jinzhao Cui, Ting Zhang, Yifeng Shen, Xiaoli Wang, Liuyang Yang, Xuefeng Huang, Qiang Huang, Yu Yang, Weizhong Yang, Zhongjie Li","doi":"10.2196/73631","DOIUrl":"10.2196/73631","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Seasonal influenza is a major global public health concern, leading to escalated morbidity and mortality rates. Traditional early warning models rely on binary (0/1) classification methods, which issue alerts only when predefined thresholds are crossed. However, these models exhibit inflexibility, often leading to false alarms or missed warnings and failing to provide granular risk assessments essential for decision-making. Therefore, we propose a probability-based early warning system using machine learning to mitigate these limitations and to offer continuous risk estimations of alerts (0-1 variable) instead of rigid threshold-based alerts. Based on probabilistic prediction, public health experts can make more flexible decisions in combination with the actual situation, significantly reducing the uncertainty and pressure in the decision-making process and reducing the waste of public health resources and the risk of social panic.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Objective: &lt;/strong&gt;The main aim of this study is to devise an innovative approach for early warning systems focused on influenza-like cases. Therefore, a Dense Residual Network (Dense ResNet), a supervised deep learning model, was developed. The model's training involved fitting the influenza-like illness positive rate, which enabled the early detection and warning of signals of changes occurring in the activity level of influenza-like cases. This departure from conventional methodologies underscores the transformative potential of machine learning, particularly in providing advanced capabilities for timely and proactive warnings in the context of influenza outbreaks.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;We developed a Dense ResNet machine learning model trained on influenza surveillance data from Northern and Southern China (2014-2024). This model generates early warning signals 3, 5, and 7 days in advance, providing a probability-based risk assessment represented as a continuous variable ranging from 0 to 1, in contrast to the traditional binary (0/1) warning systems. We evaluated the performance of this model using area under the curve scores, accuracy, recall, and F1-scores, then compared it with support vector machine (SVM), random forests, XGBoost (Extreme Gradient Boosting), and LSTM (long short-term memory) models.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;The Dense ResNet model demonstrated the best performance, characterized by 5-day lead warnings and a 50th percentile probability threshold, achieving area under the curve scores of 0.94 (Northern China) and 0.95 (Southern China). Relative to traditional models, probability-based warning signals improved early detection, reduced false alarms, and facilitated tiered public health responses.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Conclusions: &lt;/strong&gt;This study presented a novel probability-based machine learning model essential for early warning signals of influenza, demonstrating superior accuracy, flexibility, and practical applicability compared to o","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e73631"},"PeriodicalIF":3.8,"publicationDate":"2025-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12327961/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144796266","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":"AI Scribes in Health Care: Balancing Transformative Potential With Responsible Integration.","authors":"Tiffany I Leung, Andrew J Coristine, Arriel Benis","doi":"10.2196/80898","DOIUrl":"10.2196/80898","url":null,"abstract":"<p><strong>Unlabelled: </strong>The administrative burden of clinical documentation contributes to health care practitioner burnout and diverts valuable time away from direct patient care. Ambient artificial intelligence (AI) scribes-also called \"digital scribes\" or \"AI scribes\"-are emerging as a promising solution, given their potential to automate clinical note generation and reduce clinician workload, and those specifically built on a large language model (LLM) are emerging as technologies for facilitating real-time clinical documentation tasks. This potentially transformative development has a foundation on longer-standing, AI-based transcription software, which uses automated speech recognition and/or natural language processing. Recent studies have highlighted the potential impact of ambient AI scribes on clinician well-being, workflow efficiency, documentation quality, user experience, and patient interaction. So far, limited evidence indicates that ambient AI scribes are associated with reduced clinician burnout, lower cognitive task load, and significant time savings in documentation, particularly in after-hours electronic health record (EHR) work. One consistently reported benefit is the improvement in the patient-physician interaction, as physicians feel more present during a clinical encounter. However, these benefits are counterbalanced by persisting concerns regarding the accuracy, consistency, language use, and style of AI-generated notes. Studies noting errors, omissions, or hallucinations caution that diligent clinician oversight is necessary. The user experience is also heterogeneous, with benefits varying by specialty and individual workflow. Further, there are concerns about ethical and legal issues, algorithmic bias, the potential for long-term \"cognitive debt\" from overreliance on AI, and even the potential loss of physician autonomy. Additional pragmatic concerns include security, privacy, integration, interoperability, user acceptance and training, and the cost-effectiveness of adoption at scale. Finally, limited studies describe adoption or evaluation of these technologies by nonphysician clinicians and health professionals. Although ambient AI scribes and AI-driven documentation technologies are promising as potentially practice-changing tools, there are many questions remaining. Key issues persist, including responsible deployment with the goal of ensuring that ambient AI scribes produce clinical documentation that supports more efficient, equitable, and patient-centered care. To advance our collective understanding and address key issues, JMIR Medical Informatics is launching a call for papers for a new section on \"Ambient AI Scribes and AI-Driven Documentation Technologies.\" As editors, we look forward to the opportunity to advance the science and understanding of these fields through publishing high-quality and rigorous scholarly work in this new section of JMIR Medical Informatics.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e80898"},"PeriodicalIF":3.8,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12316405/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144765851","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":"Estimating Nurse Workload Using a Predictive Model From Routine Hospital Data: Algorithm Development and Validation.","authors":"Paul Meredith, Christina Saville, Chiara Dall'Ora, Tom Weeks, Sue Wierzbicki, Peter Griffiths","doi":"10.2196/71666","DOIUrl":"10.2196/71666","url":null,"abstract":"<p><strong>Background: </strong>Managing nurse staffing is complex due to fluctuating demand based on ward occupancy, patient acuity, and dependency. Monitoring staffing adequacy in real time has the potential to inform safe and efficient deployment of staff. Patient classification systems (PCSs) are being used for per shift workload measurement, but they add a frequent administrative task for ward nursing staff.</p><p><strong>Objective: </strong>The objective of this study is to explore whether an algorithm could estimate ward workload using existing routinely recorded data.</p><p><strong>Methods: </strong>Anonymized admission records and assessments from a PCS supporting the safer nursing care tool were used to determine nursing care demand in medical and surgical wards in a single UK hospital between February 2017 and February 2020. Records were linked by ward and time. The data were split into a training set (75%) and a test set (25%). We built a predictive model of ward workload (as measured by the PCS) using routinely recorded administrative data and admission National Early Warning Score. The outcome variable was ward workload derived from the patient classifications, measured as the number of whole-time equivalent (WTE) nursing staff per patient.</p><p><strong>Results: </strong>In a test set of 11,592 ward assessments from 42 wards with a mean WTE per patient of 1.64, the model's mean absolute error was 0.078, with a mean percentage error of 4.9%. A Bland-Altman plot of the differences between the predicted values and the assessment values showed 95% of them within 0.21 WTE per patient.</p><p><strong>Conclusions: </strong>Predictions of nursing workload from a relatively small number of routinely collected variables showed moderate accuracy for general wards in 1 English hospital. This demonstrates the potential for automating assessments of nurse staffing requirements from routine data, reducing time spent on this nonclinical overhead, and improving monitoring of real-time staffing pressures.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e71666"},"PeriodicalIF":3.8,"publicationDate":"2025-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12314723/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144762381","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":"Federated Analysis With Differential Privacy in Oncology Research: Longitudinal Observational Study Across Hospital Data Warehouses.","authors":"Théo Ryffel, Perrine Créquit, Maëlle Baillet, Jason Paumier, Yasmine Marfoq, Olivier Girardot, Thierry Chanet, Ronan Sy, Louise Bayssat, Julien Mazières, Vincent Vuiblet, Julien Ancel, Maxime Dewolf, François Margraff, Camille Bachot, Jacek Chmiel","doi":"10.2196/59685","DOIUrl":"10.2196/59685","url":null,"abstract":"<p><strong>Background: </strong>Federated analytics in health care allows researchers to perform statistical queries on remote datasets without access to the raw data. This method arose from the need to perform statistical analysis on larger datasets collected at multiple health care centers while avoiding regulatory, governance, and privacy issues that might arise if raw data were collected at a central location outside the health care centers. Despite some pioneering work, federated analytics is still not widely used on real-world data, and to our knowledge, no real-world study has yet combined it with other privacy-enhancing techniques such as differential privacy (DP).</p><p><strong>Objective: </strong>The first objective of this study was to deploy a federated architecture in a real-world setting. The oncology study used for this deployment compared the medical health care management of patients with metastatic non-small cell lung cancer before and after the first wave of COVID-19 pandemic. The second goal was to test DP in this real-world scenario to assess its practicality and use as a privacy-enhancing technology.</p><p><strong>Methods: </strong>A federated architecture platform was set up in the Toulouse, Reims, and Foch centers. After harmonization of the data in each center, statistical analyses were performed using DataSHIELD (Data aggregation through anonymous summary-statistics from harmonized individual-level databases), a federated analysis R library, and a new open-source DP DataSHIELD package was implemented (dsPrivacy).</p><p><strong>Results: </strong>A total of 50 patients were enrolled in the Toulouse and Reims centers and 49 in the Foch center. We have shown that DataSHIELD is a practical tool to efficiently conduct our study across all 3 centers without exposing data on a central node, once a sufficient setup has been established to configure a secure network between hospitals. All planned aggregated results were successfully generated. We also observed that DP can be implemented in practice with promising trade-offs between privacy and accuracy, and we built a library that will prove useful for future work.</p><p><strong>Conclusions: </strong>The federated architecture platform made it possible to run a multicenter study on real-world oncology data while ensuring strong privacy guarantees using differential privacy.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e59685"},"PeriodicalIF":3.8,"publicationDate":"2025-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12312987/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144762382","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":"Leveraging Interoperable Electronic Health Record (EHR) Data for Distributed Analyses in Clinical Research: Technical Implementation Report of the HELP Study.","authors":"Julia Palm, Kutaiba Saleh, André Scherag, Danny Ammon","doi":"10.2196/68171","DOIUrl":"10.2196/68171","url":null,"abstract":"<p><strong>Background: </strong>The Medical Informatics Initiative (MII) Germany established 38 data integration centers (DIC) in university hospitals to improve health care and biomedical research through the use of electronic health record (EHR) data. To showcase the value of these DIC, the HELP (Hospital-wide Electronic Medical Record Evaluated Computerized Decision Support System to Improve Outcomes of Patients with Staphylococcal Bloodstream Infection) study was initiated as a use case. This study is a clinical trial designed to assess the impact of a computerized decision support system for managing staphylococcal bacteremia.</p><p><strong>Objective: </strong>In this paper, we present the lessons learned during the use case from a technical perspective. This paper outlines the challenges encountered and solutions developed during our initial implementation of this infrastructure, providing insights applicable to other research platforms using EHR data. These insights are organized into 3 key areas: study-specific data definition and modeling, interoperable data integration and transformation, and distributed data extraction and analysis.</p><p><strong>Methods: </strong>An interdisciplinary team of clinicians, computer scientists, and statisticians created a catalog of items to identify data elements necessary for the study's evaluation and developed a domain-specific information model. DIC developed extract-transform-load pipelines to collect the disparate, site-specific EHR data and to transform it into a common data format. Health Level Seven International (HL7) Fast Healthcare Interoperability Resources (FHIR) and the MII's core dataset profiles were adopted for consistent data representation across sites. Additionally, data not present in EHRs was gathered using structured electronic case report forms. Analysis scripts were then distributed to the sites to preprocess the data locally, followed by a central analysis of the preprocessed data to generate the final overall results.</p><p><strong>Unlabelled: </strong>Our analysis revealed significant heterogeneity in data quality and implementation of interoperability standards, requiring substantial harmonization efforts. The development of analysis scripts and data extraction processes demanded multiple iterative cycles and close collaboration with local data experts. Despite these challenges, the successful implementation demonstrated the feasibility of distributed EHR analyses while highlighting the importance of thorough data quality assessment, realistic timeline planning, and multidisciplinary expertise.</p><p><strong>Conclusions: </strong>The HELP study highlights challenges and opportunities in leveraging EHR data for clinical research, particularly in the absence of mandatory data standards and resource-intensive data harmonization efforts. Despite limitations in data availability and quality, progress in digitization and interoperability frameworks offers hope for future improveme","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e68171"},"PeriodicalIF":3.8,"publicationDate":"2025-07-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12310147/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144755151","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":"Optimizing Thyroid Nodule Management With Artificial Intelligence: Multicenter Retrospective Study on Reducing Unnecessary Fine Needle Aspirations.","authors":"Jia-Hui Ni, Yun-Yun Liu, Chao Chen, Yi-Lei Shi, Xing Zhao, Xiao-Long Li, Bei-Bei Ye, Jing-Liang Hu, Li-Chao Mou, Li-Ping Sun, Hui-Jun Fu, Xiao-Xiang Zhu, Yi-Feng Zhang, Lehang Guo, Hui-Xiong Xu","doi":"10.2196/71740","DOIUrl":"10.2196/71740","url":null,"abstract":"<p><strong>Background: </strong>Most artificial intelligence (AI) models for thyroid nodules are designed to screen for malignancy to guide further interventions; however, these models have not yet been fully implemented in clinical practice.</p><p><strong>Objective: </strong>This study aimed to evaluate AI in real clinical settings for identifying potentially benign thyroid nodules initially deemed to be at risk for malignancy by radiologists, reducing unnecessary fine needle aspiration (FNA) and optimizing management.</p><p><strong>Methods: </strong>We retrospectively collected a validation cohort of thyroid nodules that had undergone FNA. These nodules were initially assessed as \"suspicious for malignancy\" by radiologists based on ultrasound features, following standard clinical practice, which prompted further FNA procedures. Ultrasound images of these nodules were re-evaluated using a deep learning-based AI system, and its diagnostic performance was assessed in terms of correct identification of benign nodules and error identification of malignant nodules. Performance metrics such as sensitivity, specificity, and the area under the receiver operating characteristic curve were calculated. In addition, a separate comparison cohort was retrospectively assembled to compare the AI system's ability to correctly identify benign thyroid nodules with that of radiologists.</p><p><strong>Results: </strong>The validation cohort comprised 4572 thyroid nodules (benign: n=3134, 68.5%; malignant: n=1438, 31.5%). AI correctly identified 2719 (86.8% among benign nodules) and reduced unnecessary FNAs from 68.5% (3134/4572) to 9.1% (415/4572). However, 123 malignant nodules (8.6% of malignant cases) were mistakenly identified as benign, with the majority of these being of low or intermediate suspicion. In the comparison cohort, AI successfully identified 81.4% (96/118) of benign nodules. It outperformed junior and senior radiologists, who identified only 40% and 55%, respectively. The area under the curve (AUC) for the AI model was 0.88 (95% CI 0.85-0.91), demonstrating a superior AUC compared with that of the junior radiologists (AUC=0.43, 95% CI 0.36-0.50; P=.002) and senior radiologists (AUC=0.63, 95% CI 0.55-0.70; P=.003).</p><p><strong>Conclusions: </strong>Compared with radiologists, AI can better serve as a \"goalkeeper\" in reducing unnecessary FNAs by identifying benign nodules that are initially assessed as malignant by radiologists. However, active surveillance is still necessary for all these nodules since a very small number of low-aggressiveness malignant nodules may be mistakenly identified.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e71740"},"PeriodicalIF":3.8,"publicationDate":"2025-07-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12310072/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144755152","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}