BMC Medical Informatics and Decision Making最新文献

{"title":"Association between blood urea nitrogen to albumin ratio and 28-day mortality in ICU patients with acute respiratory failure: a retrospective analysis of MIMIC-IV database.","authors":"Zhen Li, Jun Xie, Qian He, Chong Li","doi":"10.1186/s12911-025-03100-w","DOIUrl":"10.1186/s12911-025-03100-w","url":null,"abstract":"<p><strong>Purpose: </strong>The effect of the blood urea nitrogen-to-albumin ratio (BAR) on 28-day mortality in intensive care unit (ICU) patients with acute respiratory failure (ARF) is unknown.</p><p><strong>Methods: </strong>Patients diagnosed with ARF were screened and randomly divided into training and validation sets (7:3) on the basis of the ICD-9 and ICD-10 diagnosis codes in the Medical Information Mart for Intensive Care IV (v.2.2) database. The primary outcome was the 28-day mortality after ICU admission. The training set was categorized into the low- and high-BAR groups on the basis of the optimal BAR cutoff values for 28-day mortality determined via receiver operating characteristic analysis. The clinical significance of the BAR was evaluated by the areas under the curve (AUCs), decision curve analysis (DCA), Kaplan-Meier (K-M) survival curve, logistic regression analyses and subgroup analysis.</p><p><strong>Results: </strong>In total, 2,766 patients were included. The 28-day mortality rate was 30.2%. The AUCs and 95% confidence interval (CI) for the BAR were AUC 0.644 (95%CI, 0.618 to 0.671) in training set. Multivariate logistic regression revealed that the BAR was an independent factor affecting the prognosis of ARF in both training and validation sets. K-M curves revealed a significant difference in 28-day mortality between the low- and high-BAR groups (p < 0.001). DCA showed moderate performance. No obvious interaction was found by subgroup analysis in most subgroups.</p><p><strong>Conclusion: </strong>The present work revealed that elevated BAR was significantly associated with worse 28-day mortality in patients with any cause of ARF. It remains to be shown whether retrospective analysis of an independent cohort can confirm the high predictive value of BAR.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"253"},"PeriodicalIF":3.3,"publicationDate":"2025-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12232855/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144574865","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":"MIDAS: a technology-enabled hub-and-spoke system for the collection and dissemination of high-quality medical datasets in India.","authors":"Dibyajyoti Maity, Rohit Satish, Raghu Dharmaraju, Vijay Chandru, Rajesh Sundaresan, Harpreet Singh, Debnath Pal","doi":"10.1186/s12911-025-03092-7","DOIUrl":"10.1186/s12911-025-03092-7","url":null,"abstract":"<p><strong>Background: </strong>The need for better AI models fuels the demand for larger and larger high-quality datasets with significant diversity. Over the years, many medical imaging datasets have been published globally, but existing datasets do not contain enough samples from the population of the Indian subcontinent, leading to subpar performance of developed AI models when deployed in India. The Medical Imaging and Information Datasets (MIDAS) India initiative was launched to address this by developing standards, protocols, and policies for gathering medical imaging data nationwide.</p><p><strong>Methods: </strong>MIDAS employs a hub-and-spoke system for data collection, where each thematic hub works with a set of spokes to collect data for a specific disease or medical condition from primary, secondary, and tertiary health centers. The data gathering is guided by standard operating procedures developed from the collaborative efforts of the participating medical institutions. The annotation protocols are based on a combination of gold-standard tests and/or agreement between experts to achieve the required labeling accuracy, depending on the data type and the intended purpose of the dataset.</p><p><strong>Results: </strong>The MIDAS platform is accessible at https://midas.iisc.ac.in/ . Two datasets are already available on MIDAS, one for oral cancer and another for dural-based pathologies, for free download. Many others are under development and review. Annotated and curated data are also available under various licenses as shared by the platform partners for the registered users. The datasets use standardized ontologies for annotations at both image and pixel-level regions of interest. The annotations undergo a review process before being published and accessible for download. Standards and guidelines for creating the datasets are evolving due to the complexity of the elements involved. Challenges are steeper, especially for data originating from early or pre-onset stages of diseases, such as dysplasia in oral cancer, where the manifestation of the disease feature(s) is sometimes unclear.</p><p><strong>Conclusion: </strong>MIDAS India aims to catalyze the AI-driven transformation of healthcare by providing high-quality annotated imaging data tailored to local needs. It supports innovation, regulatory assessment, and clinical adoption of AI tools, serving as a scalable model for other countries looking to build similar data infrastructure to enhance digital healthcare delivery.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"252"},"PeriodicalIF":3.3,"publicationDate":"2025-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12232576/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144574869","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":"medAL-suite: A software solution for creating and deploying complex clinical decision support algorithms.","authors":"Ludovico Gennaro Cobuccio, Vincent Faivre, Rainer Tan, Alan Vonlanthen, Fenella Beynon, Emmanuel Barchichat, Alain Fresco, Quentin Girard, Sinan Ucak, Sylvain Schaufelberger, Ibrahim Evans Mtebene, Peter Agrea, Emmanuel Kalisa, Gillian A Levine, Martin Norris, Sabine Renggli, Alix Miauton, Lisa Cleveley, Kristina Keitel, Julien Thabard, Valérie D'Acremont, Alexandra V Kulinkina","doi":"10.1186/s12911-025-03077-6","DOIUrl":"10.1186/s12911-025-03077-6","url":null,"abstract":"<p><strong>Background: </strong>Sub-optimal healthcare quality in low-resource settings is attributed in part to poor adherence to clinical guidelines. Clinical decision support systems (CDSS) help to integrate guideline-based algorithms into logical workflows and improve adherence to evidence-based recommendations, and hence quality of care. However, the process of translating paper-based guidelines into electronic algorithmic formats is often complex, inefficient, expensive, and error-prone due to reliance on advanced software development skills and clinical knowledge.</p><p><strong>Methods: </strong>In response to these challenges, we developed open-source software called the Medical Algorithm Suite (medAL-suite), consisting of four components, with a primary goal of increasing efficiency, accuracy, and transparency of CDSS creation by giving experienced clinicians, rather than software developers, greater control over the process. At the heart of the software suite is the medAL-creator that allows clinicians to design algorithms using a code-free drag-and-drop interface. Algorithms are subsequently automatically deployed in medAL-reader to service level clinicians in health facilities. CDSS implementers use medAL-data and medAL-hub to manage configuration, versioning, and deployment.</p><p><strong>Results: </strong>Since its development, the medAL-suite has been used to digitalize complex primary care guidelines and deployed in large-scale clinical studies in Tanzania, Rwanda, Kenya, Senegal, and India, leading to notable outcomes such as the reduction of inappropriate antibiotic prescriptions and improvement in care quality. Over 300,000 pediatric outpatient consultations have been completed in Rwanda and Tanzania to date using the digital algorithm.</p><p><strong>Discussion: </strong>The medAL-suite focused on democratized development, process-centric design, point-of-care utility, touch-screen interface, low cost, and low power consumption to contribute to sustainable digital systems in low-resource settings. Important future developments and adaptations as the software evolves should emphasize interoperability and scalability, primarily via integrating CDSS functionality into electronic medical records for a streamlined user experience that supports improved service quality at the point-of-care.</p><p><strong>Clinical trial number: </strong>Not applicable.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"249"},"PeriodicalIF":3.3,"publicationDate":"2025-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12226914/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144564612","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 predictive tools and nomogram for in-hospital mortality in critically ill cancer patients: development and external validation using retrospective cohorts.","authors":"Kaier Gu, Saisai Lu","doi":"10.1186/s12911-025-03054-z","DOIUrl":"10.1186/s12911-025-03054-z","url":null,"abstract":"<p><strong>Background: </strong>The incidence of intensive care unit (ICU) admissions and the corresponding mortality rates among cancer patients are both high. However, the existing scoring systems all lack specificity. This research seeks to establish and validate a prediction model for early forecasting of in-hospital mortality in critically ill cancer patients.</p><p><strong>Methods: </strong>A retrospective analysis was conducted utilizing data from cancer patients obtained from the eICU and MIMIC-IV databases. The least absolute shrinkage and selection operator method was employed to screen predictive factors and construct six machine learning (ML) models. These models were mainly compared in terms of their predictive performance through area under the curve (AUC) and underwent external validation. Nomograms were developed using multivariate logistic regression (LR) analysis findings. The Shapley Additive exPlanations (SHAP) method was employed to explain the variables within the ML models.</p><p><strong>Results: </strong>Twelve predictive factors were chosen to develop the ML models. Among these models, the LR model and the eXtreme gradient boosting (XGB) model demonstrated the optimal efficacy. In the external validation cohort, their AUC values reached 0.751 [95% confidence interval (CI): 0.735 - 0.768] and 0.737 (95% CI: 0.720 - 0.754), respectively. Moreover, nomograms and SHAP were employed to explain the variables. Additionally, a user-friendly web-based calculator tool was created.</p><p><strong>Conclusions: </strong>The LR and XGB models were successfully developed to predict in-hospital mortality in critically ill cancer patients. Their robust predictive ability was demonstrated in the external validation cohorts. This model can assist physicians in clinical decision-making and timely intervention.</p><p><strong>Clinical trial number: </strong>Not applicable.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"251"},"PeriodicalIF":3.3,"publicationDate":"2025-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12231625/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144564611","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 narrative review of the use of PROMs and machine learning to impact value-based clinical decision-making.","authors":"Michal Pruski, Simone Willis, Kathleen Withers","doi":"10.1186/s12911-025-03083-8","DOIUrl":"10.1186/s12911-025-03083-8","url":null,"abstract":"<p><strong>Purpose: </strong>This review summarises the studies which combined Patient Reported Outcome Measures (PROMs) and Machine Learning statistical computational techniques, to predict patient post-intervention outcomes. The aim of the project was to inform those working in value-based healthcare how Machine Learning can be used with PROMs to inform clinical practice.</p><p><strong>Methods: </strong>A systematic search strategy was developed and run in six databases. The records were reviewed by a reviewer if they matched the review scope, and these decisions were scrutinised by a second reviewer.</p><p><strong>Results: </strong>82 records pertaining to 73 studies were identified. The review highlights the breadth of PROMs tools investigated, and the wide variety of Machine Learning techniques utilised across the studies. The findings suggest that there has been some success in predicting post-intervention patient outcomes. Nevertheless, there is no clear best performing Machine Learning approach to analyse this data, and while baseline PROMs scores are often a key predictor of post-intervention scores, this cannot always be assumed to be the case. Moreover, even when studies looked at similar conditions and patient groups, often different Machine Learning techniques performed best in each study.</p><p><strong>Conclusion: </strong>This review highlights that there is a potential for PROMs and Machine Learning methodology to predict patient post-intervention outcomes, but that best performing models from other previous studies cannot simply be adopted in new clinical contexts.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"250"},"PeriodicalIF":3.3,"publicationDate":"2025-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12226851/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144564610","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 and external validation of machine learning models for the early prediction of malnutrition in critically ill patients: a prospective observational study.","authors":"Yi Liu, Yehua Xu, Lixia Guo, Zhongbin Chen, Xueqin Xia, Feng Chen, Li Tang, Hua Jiang, Caixia Xie","doi":"10.1186/s12911-025-03082-9","DOIUrl":"10.1186/s12911-025-03082-9","url":null,"abstract":"<p><strong>Background: </strong>Early detection of malnutrition in critically ill patients is crucial for timely intervention and improved clinical outcomes. However, identifying individuals at risk remains challenging due to the complexity and variability of patient conditions. This study aimed to develop and externally validate machine learning models for predicting malnutrition within 24 h of intensive care unit (ICU) admission, culminating in a web-based malnutrition prediction tool for clinical decision support.</p><p><strong>Methods: </strong>A total of 1006 critically ill adult patients (aged ≥ 18 years) were included in the model development group, and 300 adult patients comprised the external validation group. The development data were partitioned into training (80%) and testing (20%) sets. Hyperparameters were optimized via 5-fold cross-validation on the training set, eliminating the need for a separate validation set while ensuring internal validation. External validation was performed on an independent group to assess generalizability. Predictors were selected using random forest recursive feature elimination; seven machine learning models-Extreme Gradient Boosting (XGBoost), random forest, decision tree, support vector machine (SVM), Gaussian naive Bayes, k-nearest neighbor (k-NN), and logistic regression-were trained and evaluated for accuracy, precision, recall, F1 score, Area Under the Receiver Operating Characteristic Curve (AUC-ROC), Area Under the Precision-Recall Curve (AUC-PR). Model interpretability was analyzed using SHapley Additive exPlanations (SHAP) to quantify feature contributions.</p><p><strong>Results: </strong>In the development phase, among 1006 patients, 34.0% had moderate malnutrition and 17.9% severe malnutrition. The XGBoost model achieved superior predictive accuracy with an accuracy of 0.90 (95% CI = 0.86-0.94), precision of 0.92 (95% CI = 0.88-0.95), recall of 0.92 (95% CI = 0.89-0.95), F1 score of 0.92 (95% CI = 0.89-0.95), AUC-ROC of 0.98 (95% CI = 0.96-0.99), and AUC-PR of 0.97 (95% CI = 0.95-0.99) on the testing set. External validation confirmed robust performance with an accuracy of 0.75 (95% CI: 0.70-0.79), precision of 0.79 (95% CI: 0.75-0.83), recall of 0.75 (95% CI: 0.70-0.79), F1 score of 0.74 (95% CI: 0.69-0.78), AUC-ROC of 0.88 (95% CI: 0.86-0.91), and AUC-PR of 0.77 (95% CI: 0.73-0.80).</p><p><strong>Conclusions: </strong>Machine learning models, particularly XGBoost, demonstrated promising performance in early malnutrition prediction in ICU settings. The resultant web-based tool offers valuable resource for clinical decision support.</p><p><strong>Trial registration: </strong>Chinese Clinical Trial Registry ChiCTR2200058286 ( https://www.chictr.org.cn/bin/project/edit? pid=248690 ). Registered 4th April 2022. Prospectively registered.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"248"},"PeriodicalIF":3.3,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12225150/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144559287","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":"Keyword-optimized template insertion for clinical note classification via prompt-based learning.","authors":"Eugenia Alleva, Isotta Landi, Leslee J Shaw, Erwin Böttinger, Ipek Ensari, Thomas J Fuchs","doi":"10.1186/s12911-025-03071-y","DOIUrl":"10.1186/s12911-025-03071-y","url":null,"abstract":"<p><strong>Background: </strong>Prompt-based learning involves the additions of prompts (i.e., templates) to the input of pre-trained large language models (PLMs) to adapt them to specific tasks with minimal training. This technique is particularly advantageous in clinical scenarios where the amount of annotated data is limited. This study aims to investigate the impact of template position on model performance and training efficiency in clinical note classification tasks using prompt-based learning, especially in zero- and few-shot settings.</p><p><strong>Methods: </strong>We developed a keyword-optimized template insertion method (KOTI) to enhance model performance by strategically placing prompt templates near relevant clinical information within the notes. The method involves defining task-specific keywords, identifying sentences containing these keywords, and inserting the prompt template in their vicinity. We compared KOTI with standard template insertion (STI) methods in which the template is directly appended at the end of the input text. Specifically, we compared STI with naïve tail-truncation (STI-s) and STI with keyword-optimized input truncation (STI-k). Experiments were conducted using two pre-trained encoder models, GatorTron and ClinicalBERT, and two decoder models, BioGPT and ClinicalT5, across five classification tasks, including dysmenorrhea, peripheral vascular disease, depression, osteoarthritis, and smoking status classification.</p><p><strong>Results: </strong>Our experiments revealed that the KOTI approach consistently outperformed both STI-s and STI-k in zero-shot and few-shot scenarios for encoder models, with KOTI yielding a significant 24% F1 improvement over STI-k for GatorTron and 8% for Clinical BERT. Additionally, training with balanced examples further enhanced performance, particularly under few-shot conditions. In contrast, decoder-based models exhibited inconsistent results, with KOTI showing significant improvement in F1 score over STI-k for BioGPT (+19%), but a significant drop for ClinicalT5 (-18%), suggesting that KOTI is not beneficial across all transformer model architectures.</p><p><strong>Conclusion: </strong>Our findings underscore the significance of template position in prompt-based fine-tuning of encoder models and highlights KOTI's potential to optimize real-world clinical note classification tasks with few training examples.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"247"},"PeriodicalIF":3.3,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12224782/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144559288","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":"Clinicopathological observation and prognostic factors of invasive breast cancer with medullary in Xinjiang.","authors":"Yan Li, Xiaoping Ma, Zhenhui Zhao, Li Li, Chunyan Gao, Dan Liu, Bingyu Li, Bing Zhao","doi":"10.1186/s12911-025-03089-2","DOIUrl":"10.1186/s12911-025-03089-2","url":null,"abstract":"<p><strong>Background: </strong>Invasive breast cancer (BC) with medullary is a special type of BC, accounting for 3 ~ 6% of invasive BC. To investigate clinicopathological characteristics concerning invasive BC with medullary in Xinjiang area and analyze the prognostic factors of patients.</p><p><strong>Materials and methods: </strong>Clinicopathological data of100 invasive BC patients with medullary admitted to Xinjiang tumor Hospital from January 2011 to December 2020 were analyzed retrospectively via R packages. Count data were analyzed applying COX regression model and relevant statistics approaches.</p><p><strong>Results: </strong>All 100patients that investigated were female and the median age was 48 years old. The median follow-up time was 92 months. The the cumulative disease free survival (DFS) at 10 years were 84.1%. Log-Rank test results suggest that ethnicity, TNM stage, tumor size, lymph node metastasis, and pathological type are possible adverse factors affecting patients with invasive breast cancer with myeloid features (P<0.05).</p><p><strong>Conclusion: </strong>Despite the poor pathological features of Invasive BC with medullary, its clinical outcome is more favorable. Ethnicity, TNM stage, tumor size, lymph node metastasis, and pathological type are possible adverse factors affecting patients with invasive breast cancer with medullary features.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"246"},"PeriodicalIF":3.3,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12224657/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144552310","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":"Integrating an AI platform into clinical IT: BPMN processes for clinical AI model development.","authors":"Kfeel Arshad, Saman Ardalan, Björn Schreiweis, Björn Bergh","doi":"10.1186/s12911-025-03087-4","DOIUrl":"10.1186/s12911-025-03087-4","url":null,"abstract":"<p><strong>Background: </strong>There has been a resurgence of Artificial Intelligence (AI) on a global scale in recent times, resulting in the development of cutting-edge AI solutions within hospitals. However, this has also led to the creation of isolated AI solutions that are not integrated into clinical IT. To tackle this issue, a clinical Artificial Intelligence (AI) platform that handles the entire development cycle of clinical AI models and is integrated into clinical IT is required. This research investigates the integration of a clinical AI platform into the clinical IT infrastructure. This is demonstrated by outlining the stages of the AI model development cycle within the clinical IT infrastructure, illustrating the interaction between different IT system landscapes within the hospital with BPMN diagrams.</p><p><strong>Methods: </strong>Initially, a thorough analysis of the requirements is conducted to refine the necessary aspects of the clinical AI platform with consideration of the individual aspects of clinical IT. Subsequently, processes representing the entire development cycle of an AI model are identified. To facilitate the architecture of the AI platform, BPMN diagrams of all the identified processes are created. Clinical use cases are used to evaluate the processes using the FEDS framework.</p><p><strong>Results: </strong>Our BPMN process diagrams cover the entire development cycle of a clinical AI model within the clinical IT. The processes involved are Data Selection, Data Annotation, On-site Training and Testing, and Inference, with distinctions between (Semi-Automated) Batch Inference and Real-Time Inference. Three clinical use cases were assessed to evaluate the processes and demonstrate that this approach covers a wide range of clinical AI use cases.</p><p><strong>Conclusions: </strong>The evaluations were executed successfully, which indicate the comprehensive nature of our approach. The results have shown that different clinical AI use cases are covered by the BPMN diagrams. Our clinical AI platform is ideally suited for the local development of AI models within clinical IT. This approach provides a basis for further developments, e.g., enabling the training and deployment of an AI model across multiple sites or the integration of security- and privacy-related aspects.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"243"},"PeriodicalIF":3.3,"publicationDate":"2025-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12218938/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144552312","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":"Exploring professionals' experiences with secure messaging in Dutch outpatient clinics: emerging differences in use frequencies and types across medical specialties.","authors":"Marjolein A G van Offenbeek, Oskar P Roemeling, Anne Burggraaf","doi":"10.1186/s12911-025-03081-w","DOIUrl":"10.1186/s12911-025-03081-w","url":null,"abstract":"<p><strong>Background: </strong>Secure digital messaging is a two-way communication channel that gained ground in healthcare over the past decade. While a direct channel between patients and providers may support patients, professionals' work pressure makes it imperative that patient-provider communication remains efficient. Thus far, there is little insight into how the use of digital messaging between outpatients and professionals varies across medical specialties and how professionals experience effects on their workload and patient empowerment.</p><p><strong>Methods: </strong>We conducted a two-stage, cross-specialty study in a Dutch hospital. Stage one analyzed differences in outpatient clinics' (n = 25) messaging frequencies over a 16-month period. In stage two, across seven outpatient clinics, purposively selected to maximize variation, we interviewed 15 professionals and coded these data for use types and professionals' experiences, followed by a focus group to check the findings.</p><p><strong>Results: </strong>While overall use increased, use frequencies varied across specialties from 228 to 31,319 over the 16-month period. The number of messages per patient ranged between 1 and 274. Eight patient-provider use types emerged: asking and answering administrative questions, asking and answering medical questions, medical updates, sending out information, enquiries about patient updates, and social updates. Most use types were experienced as partial substitutes for phone calls, emails, or both. Only social updates were seen to constitute a complementary form of patient-driven communication. Professionals experienced messaging as inefficient when synchronicity was required and for acute questions. For chronic patient streams of internal medicine specialties, higher frequencies and more use types were reported and greater usefulness was experienced than for surgical patient streams, which was unrelated to patient numbers. The interviewed professionals felt that messaging empowered patients but increased their workload. This increase depended on how messaging use was coordinated and patient expectations managed.</p><p><strong>Conclusions: </strong>Professionals may welcome messaging for patient empowerment, but in our study did so less for substitution-based efficiency effects. In chronic care settings-where communication between patients and providers is seen as integral to care delivery-messaging may be valued despite the potential for increased workload. In contrast, in surgical settings, messaging may be viewed as an additional, non-reimbursable service rather than a core care component.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"245"},"PeriodicalIF":3.3,"publicationDate":"2025-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12225041/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144552311","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}