JMIR Medical Informatics最新文献

{"title":"Author's Reply: \"Data Contamination in AI Evaluation\".","authors":"ChulHyoung Park, Min Ho An, Gyubeom Hwang, Rae Woong Park, Juho An","doi":"10.2196/82057","DOIUrl":"https://doi.org/10.2196/82057","url":null,"abstract":"","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e82057"},"PeriodicalIF":3.8,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145187591","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predicting Waiting Times for Medical Tasks in a Pediatric Hospital Using Machine Learning: Comprehensive, Retrospective, Real-World Study.","authors":"Lin Lin Guo, Rui Tang, Jia Yang Wang, Si Zheng, Yin Zeng, Jun Hou, Mo Chen Dong, Jiao Li, Ying Cui","doi":"10.2196/77297","DOIUrl":"10.2196/77297","url":null,"abstract":"<p><strong>Background: </strong>The shortage of pediatric medical resources and overcrowding in children's hospitals are severe issues in China. Accurately predicting waiting times can help optimize hospital operational efficiency.</p><p><strong>Objective: </strong>This study aims to develop machine learning models to predict waiting times for various laboratory and radiology examinations at a pediatric hospital.</p><p><strong>Methods: </strong>Time stamp data from laboratory and radiology examinations were retrospectively collected from the pediatric hospital information system between November 1, 2024, and March 13, 2025. Two queue-related and 4 time-based features were extracted using queue theory. Linear regression and 8 machine learning models were trained to predict waiting times for each medical task. Hyperparameters were fine-tuned using randomized search and 10-fold cross-validation, and the bootstrap method was used for model evaluation. Mean absolute error, mean square error, root mean square error, and the coefficient of determination (R²) were used as evaluation metrics. Shapley additive explanations values were used to assess feature importance.</p><p><strong>Results: </strong>A total of 230,864 time-stamped records were included after data preprocessing. The median waiting time was 4.817 (IQR 1.867-12.050) minutes for all medical tasks. Waiting times for radiology examinations were generally longer than those for laboratory tests. Tree-based algorithms, such as random forest and classification and regression trees, performed best in predicting laboratory test waiting times, with R² values ranging from mean 0.880(SD 0.003) to mean 0.934 (SD 0.003). However, the machine learning models did not perform well in predicting radiology examination waiting times, with R2 ranging from 0.114 (SD 0.005) to 0.719 (SD 0.004). Feature importance analysis revealed that queue-related predictors, especially the number of queuing patients, were the most important in predicting waiting times.</p><p><strong>Conclusions: </strong>Task-specific prediction models are more appropriate for accurately predicting waiting times across various medical tasks. Guided by queue theory principles, we developed machine learning models for the waiting time prediction of each medical task and highlighted the importance of queue-related predictors.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e77297"},"PeriodicalIF":3.8,"publicationDate":"2025-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12481172/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145194016","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":"Building a Standardized Cancer Synoptic Report With Semantic and Syntactic Interoperability: Development Study Using SNOMED CT and Fast Healthcare Interoperability Resources (FHIR).","authors":"Jieun Hwang, Alexander K Goel, Brian A Rous, George Birdsong, Paul A Seegers, Stefan Dubois, Thomas Rüdiger, Walter S Campbell","doi":"10.2196/76870","DOIUrl":"10.2196/76870","url":null,"abstract":"<p><strong>Background: </strong>Pathology reports contain critical information necessary for the management of cancer patient care. Efforts to structure pathology cancer reports by the College of American Pathologists and the International Collaboration on Cancer Reporting (ICCR) have been successful in standardizing pathology reports. Likewise, standards development organizations have advanced methods to improve data computability and exchange, by enabling interoperability of pathology cancer reports.</p><p><strong>Objective: </strong>This study aimed to provide a tractable method to render pathology cancer reports computable and interoperable using published cancer reporting protocols, SNOMED Clinical Terms (SNOMED CT) and Health Level 7 (HL7) Fast Healthcare Interoperability Resources (FHIR).</p><p><strong>Methods: </strong>The ICCR colorectal cancer (CRC) reporting dataset (version 1.0) was evaluated by terminologists and pathologists. SNOMED CT concepts were bound to the data elements. The dataset was then converted into a FHIR structured data capture (SDC) questionnaire using the United States National Library of Medicine tooling and rendered into a FHIR-conformant message for data exchange.</p><p><strong>Results: </strong>The ICCR CRC dataset contained 216 data elements; 207 data elements were bound to SNOMED CT and incorporated into a FHIR SDC construct. The 9 uncoded data elements were ambiguous and could not be reliably encoded. The resultant FHIR SDC form fully represented the ICCR CRC dataset and rendered these data in an R4 JSON format for data exchange.</p><p><strong>Conclusions: </strong>This study demonstrates a tractable and extensible approach to making cancer pathology reports fully computable and interoperable that can be broadly adopted. ICCR datasets are supported internationally and supported by multiple national pathology societies. These datasets can be fully represented using SNOMED CT to render data elements computable and semantically faithful to their intended meaning. The use of the FHIR SDC construct enables widespread and standardized data exchange of clinical information. While challenges remain, including FHIR adoption and the need to maintain current clinical content and standard terminology, this approach provides a clear pathway toward implementation.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e76870"},"PeriodicalIF":3.8,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12463337/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145151561","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":"Interpretable Machine Learning Model for Pulmonary Hypertension Risk Prediction: Retrospective Cohort Study.","authors":"Hongxia Jiang, Han Gao, Dexin Wang, Qingli Zeng, Xiaojun Hao, Zhenshun Cheng","doi":"10.2196/74117","DOIUrl":"10.2196/74117","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Pulmonary hypertension (PH) is a progressive disorder characterized by elevated pulmonary artery pressure and increased pulmonary vascular resistance, ultimately leading to right heart failure. Early detection is critical for improving patient outcomes.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Objective: &lt;/strong&gt;The diagnosis of PH primarily relies on right heart catheterization, but its invasive nature significantly limits its clinical use. Echocardiography, as the most common noninvasive screening and diagnostic tool for PH, provides valuable patient information. This study aims to identify key PH predictors from echocardiographic parameters, laboratory tests, and demographic data using machine learning, ultimately constructing a predictive model to support early noninvasive diagnosis of PH.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;This study compiled comprehensive datasets comprising echocardiography measurements, clinical laboratory data, and fundamental demographic information from patients with PH and matched controls. The final analytical cohort consisted of 895 participants with 85 evaluated variables. Recursive feature elimination was used to select the most relevant echocardiographic variables, which were subsequently integrated into a composite ultrasound index using machine learning techniques, XGBoost (Extreme Gradient Boosting). LASSO (least absolute shrinkage and selection operator) regression was applied to select the potential predictive variable from laboratory tests. Then, the ultrasound index variables and selected laboratory tests were combined to construct a logistic regression model for the predictive diagnosis of PH. The model's performance was rigorously evaluated using receiver operating characteristic curves, calibration plots, and decision curve analysis to ensure its clinical relevance and accuracy. Both internal and external validation were used to assess the performance of the constructed model.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;A total of 16 echocardiographic parameters (right atrium diameter, pulmonary artery diameter, left atrium diameter, tricuspid valve reflux degree, right ventricular diameter, E/E' [ratio of mitral valve early diastolic inflow velocity (E) to mitral annulus early diastolic velocity (E')], interventricular septal thickness, left ventricular diameter, ascending aortic diameter, left ventricular ejection fraction, left ventricular outflow tract velocity, mitral valve reflux degree, pulmonary valve outflow velocity, mitral valve inflow velocity, aortic valve reflux degree, and left ventricular posterior wall thickness) combined with 2 laboratory biomarkers (prothrombin time activity and cystatin C) were identified as optimal predictors, forming a high-performance PH prediction model. The diagnostic model demonstrated high predictive accuracy, with an area under the receiver operating characteristic curve of 0.997 in the internal validation and 0.974 in the external validation. Both calibration plots a","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e74117"},"PeriodicalIF":3.8,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12459742/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145139688","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":"Predicting Lymph Node Metastasis in Rectal Cancer: Development and Validation of a Machine Learning Model Using Clinical Data.","authors":"Wei Hou, Chuangwei Li, Zhen Wang, Wanqin Wang, Shouhong Wan, Bingbing Zou","doi":"10.2196/73765","DOIUrl":"10.2196/73765","url":null,"abstract":"<p><strong>Background: </strong>Rectal cancer (RC) is a common malignant tumor, with lymph node metastasis (LNM) being a critical determinant of patient prognosis. Traditional diagnostic methods have limitations, necessitating the development of predictive models using clinical data.</p><p><strong>Objective: </strong>This study aimed to construct and validate machine learning (ML) models to predict LNM risk in patients with RC based on clinical data.</p><p><strong>Methods: </strong>Retrospective data from 2454 patients with RC (SEER [Surveillance, Epidemiology, and End Results] database) were split into training (n=1954) and internal validation (n=500) sets. An external cohort (n=500) was obtained from the First Affiliated Hospital of Anhui Medical University. Lymph node features identified via computed tomographic scans were integrated with clinicopathological data. Variables were selected using LASSO (Least Absolute Shrinkage and Selection Operator), followed by univariate and multivariate logistic regression. Eleven ML models (Logistic Regression, K-Nearest Neighbors, Extremely Randomized Trees, Naive Bayes, XGBoost [XBG], Light Gradient Boosting Machine, Multilayer Perceptron, Gradient Boosting, Support Vector Machine, Random Forest, and Ada-Boost) were evaluated via area under the receiver operating characteristic curve (AUC), calibration curves, and decision curve analysis.</p><p><strong>Results: </strong>LNM prevalence was 26.9% (training), 27% (internal validation), and 81% (external validation). Independent LNM predictors included tumor grade, clinical T stage, N stage, tumor length, neural invasion, and total lymph nodes. Internal validation AUC ranged from 0.859 to 0.964; external validation AUC was 0.735-0.838. In the internal validation set, Random Forest and Extremely Randomized Trees achieved the highest AUC (0.964, 95% CI 0.950-0.978), while XGB demonstrated superior cross-cohort stability (AUC 0.942, 95% CI 0.925-0.959). For external validation, Gradient Boosting had the highest AUC (0.838, 95% CI 0.801-0.875), followed by XGB (0.832, 95%CI 0.794-0.869). XGB showed minimal calibration error with curves closest to the ideal diagonal and yielded the highest net benefit in decision curve analysis across critical thresholds.</p><p><strong>Conclusions: </strong>This study successfully developed and validated 11 ML models to predict LNM risk in RC. The XGB model was optimal, achieving an AUC >0.9 in 10 internal models and an AUC >0.8 in 7 external models. The identified predictors of LNM can facilitate early diagnosis and personalized treatment, highlighting the potential of integrating computed tomographic scan data with clinicopathological findings to build effective predictive models.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e73765"},"PeriodicalIF":3.8,"publicationDate":"2025-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12456929/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145132880","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":"Virtual Transition of Care Clinics and Associated Readmission Rates: 3-Year Retrospective Cohort Study.","authors":"Sarah F Horman, Milla Kviatkovsky, Edward Castillo, Patricia Maysent, Chad VanDenBerg, John Bell, Christopher A Longhurst","doi":"10.2196/73495","DOIUrl":"https://doi.org/10.2196/73495","url":null,"abstract":"<p><strong>Background: </strong>Hospital readmissions pose a significant burden on patients, health care providers, and systems, with an estimated annual cost of $17 billion. Timely follow-up within 7 days postdischarge is known to reduce readmissions but is often limited by access constraints. While transitions of care clinics have demonstrated benefits in reducing unplanned readmissions, physical space requirements can be logistically and financially challenging.</p><p><strong>Objective: </strong>This study aimed to evaluate the effectiveness of a virtual transitions of care (VToC) clinic in reducing 30-day hospital readmissions and improving postdischarge care coordination.</p><p><strong>Methods: </strong>University of California, San Diego Health implemented a hospitalist-led VToC clinic designed to support clinical management, medication reconciliation, primary care provider repatriation, and specialty care navigation. The study included 2314 patients seen in the VToC clinic between September 2021 and September 2024. Outcomes were compared to a benchmark group using regression analysis to assess the impact on 30-day readmission rates.</p><p><strong>Results: </strong>The 30-day readmission rate for VToC patients was 14.9% (344/2314), significantly lower than the 20.1% (4659/23,129) observed in the benchmark group (P<.001). Regression analysis indicated that patients not participating in the VToC clinic had a higher likelihood of readmission (odds ratio=1.37; 95% CI=1.21-1.54; P<.001). The most substantial reduction in readmissions was observed among patients with moderate readmission risk (LACE+ score of 50-75).</p><p><strong>Conclusions: </strong>VToC clinics are a feasible and effective strategy for enhancing postdischarge care, reducing hospital readmissions, and improving care coordination. This model supports the quadruple aim by promoting better health outcomes, improved patient experience, cost-efficiency, and care equity.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e73495"},"PeriodicalIF":3.8,"publicationDate":"2025-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145132940","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Machine Learning and Shapley Additive Explanations Value Integration for Predicting the Prognostic of Anti-N-Methyl-D-Aspartate Receptor Encephalitis: Model Development and Evaluation Study.","authors":"Jia Wang, Haotian Wu, Han Cai, YingXiang Wang, Jian Gu","doi":"10.2196/75020","DOIUrl":"10.2196/75020","url":null,"abstract":"<p><strong>Background: </strong>Anti-N-methyl-D-aspartate receptor (NMDAR) encephalitis is a rare disease with no accurate prognostic tools to predict the prognosis of patients.</p><p><strong>Objective: </strong>This study aims to develop an interpretable machine learning model using real-world clinical data to guide personalized therapeutic strategies.</p><p><strong>Methods: </strong>This retrospective cohort study analyzed 140 patients with NMDAR encephalitis treated at the Third Affiliated Hospital of Sun Yat-sen University (2015-2024). Feature selection was done using recursive feature elimination. The model was constructed by 3 machine learning algorithms: decision tree, random forest (RF), and extreme gradient boosting. Mean squared error, root-mean-squared error, R² (coefficient of determination), mean absolute error, and mean absolute percentage error were used to evaluate the model performance. Finally, the optimal model was interpreted via Shapley Additive Explanations (SHAP) and deployed as a web application using the Flask framework.</p><p><strong>Results: </strong>The median age of patients with anti-NMDAR encephalitis was 23 (IQR 18-31.8) years. The median Clinical Evaluation Scale for Autoimmune Encephalitis score at acute onset was 11 (IQR 6-16). After preprocessing, 20 features, including 4 demographic characteristics, 3 clinical characteristics, 11 laboratory parameters, and 2 neuroimaging characteristics, were selected. The RF demonstrated superior accuracy in predicting the prognosis (mean squared error=11.01; root-mean-squared error=3.32; R²=0.71; mean absolute error=2.49; mean absolute percentage error=0.48). SHAP analysis identified admission to the intensive care unit (mean |SHAP value|=1.65), initial symptoms-memory deficits (0.69), and uric acid (0.53) as the most important prognostic predictors.</p><p><strong>Conclusions: </strong>We developed and validated an interpretable RF-based prognostic model for NMDAR encephalitis. The web-deployed tool enables real-time risk stratification, facilitating clinical decision-making and personalized therapeutic interventions for clinicians.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e75020"},"PeriodicalIF":3.8,"publicationDate":"2025-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12453450/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145126417","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":"Home Monitoring Delivered Through the Emergency Department for Outpatients With COVID-19: COVID19@Home Aachen Pilot Cohort Study.","authors":"Lukas Niekrenz, Christian Hübel, Christopher Plata, Henning Biermann, Claas Leber, Lisa Sophie Schütze, Svea Holtz, Susanne Maria Köhler, Kim Deutsch, Nurlan Dauletbayev, Sebastian Kuhn, Beate Sigrid Müller, Christian Cornelissen, Rembert Koczulla, Gernot Rohde, Claus Franz Vogelmeier, Jörg Christian Brokmann, Michael Dreher","doi":"10.2196/58364","DOIUrl":"10.2196/58364","url":null,"abstract":"<p><strong>Background: </strong>The overwhelming COVID-19 situation in 2020/2021 required novel approaches that did not require additional personnel within the current health care system. Therefore, we initiated a trial of nonsupervised home monitoring via the emergency department of a tertiary hospital without the support of a virtual ward as part of the \"Netzwerk Universitaetsmedizin\" cooperation in Germany. Given that daily vital sign checks for inpatients with COVID-19 could indicate clinical deterioration, this approach might also be helpful in an outpatient setting and could help to identify the need for hospitalization and additional resources.</p><p><strong>Objective: </strong>This study aims to determine whether patient-led home monitoring for acute SARS-CoV-2 infection can be implemented through the emergency department of a tertiary care provider.</p><p><strong>Methods: </strong>Patients who tested positive for SARS-CoV-2 infection in our emergency department between May 2021 and May 2022, did not have a medical indication for hospitalization, and were discharged to the outpatient setting were offered the opportunity to perform nonsupervised home monitoring of vital signs. Those who agreed to participate received Bluetooth-enabled devices to measure temperature, oxygen saturation, and blood pressure and downloaded a smartphone app. Participants were encouraged to measure their vital signs for at least 28 days. There was no virtual ward or real-time surveillance of the recorded data, but these could be made available to primary care physicians. Compliance with self-measurements was determined, and participants were contacted after the monitoring period for a semistructured interview.</p><p><strong>Results: </strong>A total of 828 patients with COVID-19 were treated at the emergency department. Of these, 262 were directly discharged into ambulatory isolation after initial assessment, 25 were offered the opportunity of nonsupervised home monitoring, 15 successfully activated the devices, and 9 performed more than one complete measurement using the app. These 9 participants used the devices for an average of 15.8 days after discharge. Interviewed participants reported various difficulties with device setup but said they were pleased to use home monitoring and felt that the measurement option gave them additional security.</p><p><strong>Conclusions: </strong>This study highlighted the challenges associated with implementing nonsupervised home monitoring for outpatients with COVID-19 who presented to the emergency department of a tertiary hospital. Implementing such a system without the involvement of additional personnel does not appear to be the optimal approach. We suggest that the physician-patient relationship might be a factor that is essential for the success of patient-led approaches to home monitoring.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e58364"},"PeriodicalIF":3.8,"publicationDate":"2025-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12445781/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145088433","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":"Development of a New Electronic Death Certificate and Death Management Module Integrated Into the Health Information System of a Tertiary Hospital in Mali: Implementation Report.","authors":"Mahmoud Cissé, Mahamoudane Niang, Abdrahamane Anne, Mariam Sidibé, Dramane Traoré, Abdoulaye Traoré, Idrissa Traoré, Cheick Oumar Bagayoko","doi":"10.2196/62949","DOIUrl":"10.2196/62949","url":null,"abstract":"<p><strong>Background: </strong>Death certification provides reliable epidemiological data that are essential for public health decision-making. Implementing an electronic death certificate improves data quality, accuracy, and timeliness. Recognizing its importance, we developed and integrated a mortality management module into the hospital \"Le Luxembourg,\" enhancing mortality data collection and utilization.</p><p><strong>Objective: </strong>This implementation aimed to improve the completeness and accuracy of mortality data, accelerate the timeliness of death reporting, and facilitate the downstream use of data to inform public health planning and research.</p><p><strong>Methods: </strong>We began by analyzing the existing infrastructure and organizational setup within the hospital. Through a series of interviews, we identified the needs of all users, enabling the design of a module suited to all levels of operation. We implemented a module comprising multiple functionalities, including certificate editing, validation, storage, mortality statistics generation. It also integrates ICD-10 coding and follows the World Health Organization (WHO) model, while remaining adapted to the hospital's specific context. To ensure optimal usability, we assembled a project team that included a mortality audit committee. After implementation, all users received training and continuous direct technical support was provided.</p><p><strong>Results: </strong>We developed a new death certificate model in line with WHO recommendations. Access to the certificate is secured by a unique username and password. To improve data quality, the certification process involves several validation steps: initial recording, which can be modified when the medical section is not completed by a senior physician; pre-validation by the senior physician and final validation by the mortality audit committee. The chain of morbid events is documented using ICD-10 diagnoses. Beyond the certificate itself, the system also allows for civil registration of the death. Moreover, the module can generate statistics based on multiple criteria. This process takes place with the involvement and active engagement of all stakeholders.</p><p><strong>Conclusions: </strong>We established a unique, secure, WHO-compliant death certificate model that ensures high-quality, easily exploitable, and well-archived data. The experience of this hospital may serve as a foundation for scaling up this model to other healthcare facilities within the country.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e62949"},"PeriodicalIF":3.8,"publicationDate":"2025-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12445778/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145088428","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":"Fusion of X-Ray Images and Clinical Data for a Multimodal Deep Learning Prediction Model of Osteoporosis: Algorithm Development and Validation Study.","authors":"Jun Tang, Xiang Yin, Jiangyuan Lai, Keyu Luo, Dongdong Wu","doi":"10.2196/70738","DOIUrl":"10.2196/70738","url":null,"abstract":"<p><strong>Background: </strong>Osteoporosis is a bone disease characterized by reduced bone mineral density and mass, which increase the risk of fragility fractures in patients. Artificial intelligence can mine imaging features specific to different bone densities, shapes, and structures and fuse other multimodal features for synergistic diagnosis to improve prediction accuracy.</p><p><strong>Objective: </strong>This study aims to develop a multimodal model that fuses chest X-rays and clinical parameters for opportunistic screening of osteoporosis and to compare and analyze the experimental results with existing methods.</p><p><strong>Methods: </strong>We used multimodal data, including chest X-ray images and clinical data, from a total of 1780 patients at Chongqing Daping Hospital from January 2019 to August 2024. We adopted a probability fusion strategy to construct a multimodal model. In our model, we used a convolutional neural network as the backbone network for image processing and fine-tuned it using a transfer learning technique to suit the specific task of this study. In addition, we introduced a gradient-based wavelet feature extraction method. We combined it with an attention mechanism to assist in feature fusion, which enhanced the model's focus on key regions of the image and further improved its ability to extract image features.</p><p><strong>Results: </strong>The multimodal model proposed in this paper outperforms the traditional methods in the 4 evaluation metrics of area under the curve value, accuracy, sensitivity, and specificity. Compared with using only the X-ray image model, the multimodal model improved the area under the curve value significantly from 0.951 to 0.975 (P=.004), the accuracy from 89.32% to 92.36% (P=.045), the sensitivity from 89.82% to 91.23% (P=.03), and the specificity from 88.64% to 93.92% (P=.008).</p><p><strong>Conclusions: </strong>While the multimodal model that fuses chest X-ray images and clinical data demonstrated superior performance compared to unimodal models and traditional methods, this study has several limitations. The dataset size may not be sufficient to capture the full diversity of the population. The retrospective nature of the study may introduce selection bias, and the lack of external validation limits the generalizability of the findings. Future studies should address these limitations by incorporating larger, more diverse datasets and conducting rigorous external validation to further establish the model's clinical use.</p>","PeriodicalId":56334,"journal":{"name":"JMIR Medical Informatics","volume":"13 ","pages":"e70738"},"PeriodicalIF":3.8,"publicationDate":"2025-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12445622/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145088379","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}