International Journal of Medical Informatics最新文献

{"title":"Online professionalism through the lens of medical students and residents: A focus group study","authors":"Sebastiaan A. Pronk ,&nbsp;Simone L. Gorter ,&nbsp;Scheltus J. van Luijk ,&nbsp;Guy J. Oudhuis ,&nbsp;Pieter C. Barnhoorn ,&nbsp;Walther N.K.A. van Mook","doi":"10.1016/j.ijmedinf.2025.105879","DOIUrl":"10.1016/j.ijmedinf.2025.105879","url":null,"abstract":"<div><h3>Purpose</h3><div>Social media influences the practice of healthcare professionals. Existing studies on online professionalism and social media are scarce, and most used survey-based methods. This qualitative study explores online professionalism in healthcare among medical students and residents and maps their perceived educational needs.</div></div><div><h3>Method</h3><div>Semi-structured focus group interviews were conducted between September 2019 and June 2021 to explore the perceptions of online professionalism among Dutch medical students and residents. Interviews were recorded, transcribed, and thematically analyzed iteratively and independently by two researchers.</div></div><div><h3>Results</h3><div>Seven one-hour focus groups were conducted among 24 medical students and 22 residents. Patient requests from social media directed at students and residents occurred, none were accepted. Upon patient confidentiality breaches on social media, peers tended to speak up to one another. Participants voiced that clarity about the ‘grey areas’ − where distinguishing between right and wrong is difficult − of social media use is needed.</div></div><div><h3>Conclusions</h3><div>Social media use was widespread among participants and patients’ requests directed at students and residents did occur. They were unlikely to speak about online professionalism lapses to their peers unless a breach of patient confidentiality is involved. Educators should focus on enhancing the professional use of social media in both undergraduate and postgraduate training.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"198 ","pages":"Article 105879"},"PeriodicalIF":3.7,"publicationDate":"2025-03-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143629194","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluating AI-generated patient education materials for spinal surgeries: Comparative analysis of readability and DISCERN quality across ChatGPT and deepseek models","authors":"Mi Zhou ,&nbsp;Yun Pan ,&nbsp;Yuye Zhang ,&nbsp;Xiaomei Song ,&nbsp;Youbin Zhou","doi":"10.1016/j.ijmedinf.2025.105871","DOIUrl":"10.1016/j.ijmedinf.2025.105871","url":null,"abstract":"<div><h3>Background</h3><div>Access to patient-centered health information is essential for informed decision-making. However, online medical resources vary in quality and often fail to accommodate differing degrees of health literacy. This issue is particularly evident in surgical contexts, where complex terminology obstructs patient comprehension. With the increasing reliance on AI models for supplementary medical information, the reliability and readability of AI-generated content require thorough evaluation.</div></div><div><h3>Objective</h3><div>This study aimed to evaluate four natural language processing models—ChatGPT-4o, ChatGPT-o3 mini, DeepSeek-V3, and DeepSeek-R1—in generating patient education materials for three common spinal surgeries: lumbar discectomy, spinal fusion, and decompressive laminectomy. Information quality was evaluated using the DISCERN score, and readability was assessed through Flesch-Kincaid indices.</div></div><div><h3>Results</h3><div>DeepSeek-R1 produced the most readable responses, with Flesch-Kincaid Grade Level (FKGL) scores ranging from 7.2 to 9.0, succeeded by ChatGPT-4o. In contrast, ChatGPT-o3 exhibited the lowest readability (FKGL &gt; 10.4). The DISCERN scores for all AI models were below 60, classifying the information quality as “fair,” primarily due to insufficient cited references.</div></div><div><h3>Conclusion</h3><div>All models achieved merely a “fair” quality rating, underscoring the necessity for improvements in citation practices, and personalization. Nonetheless, DeepSeek-R1 and ChatGPT-4o generated more readable surgical information than ChatGPT-o3. Given that enhanced readability can improve patient engagement, reduce anxiety, and contribute to better surgical outcomes, these two models should be prioritized for assisting patients in the clinical.</div></div><div><h3>Limitation &amp; Future direction</h3><div>This study is limited by the rapid evolution of AI models, its exclusive focus on spinal surgery education, and the absence of real-world patient feedback, which may affect the generalizability and long-term applicability of the findings. Future research ought to explore interactive, multimodal approaches and incorporate patient feedback to ensure that AI-generated health information is accurate, accessible, and facilitates informed healthcare decisions.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"198 ","pages":"Article 105871"},"PeriodicalIF":3.7,"publicationDate":"2025-03-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143642842","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"AI-based personalized real-time risk prediction for behavioral management in psychiatric wards using multimodal data","authors":"Ri-Ra Kang ,&nbsp;Yong-gyom Kim ,&nbsp;Minseok Hong ,&nbsp;Yong Min Ahn ,&nbsp;KangYoon Lee","doi":"10.1016/j.ijmedinf.2025.105870","DOIUrl":"10.1016/j.ijmedinf.2025.105870","url":null,"abstract":"<div><h3>Background</h3><div>Suicide is a major global health issue, with approximately 700,000 deaths annually (WHO). In psychiatric wards, managing harmful behaviors such as suicide, self-harm, and aggression is essential to ensure patient and staff safety. However, psychiatric wards in South Korea face challenges due to high patient-to-psychiatrist ratios and heavy workloads. Current models relying on demographic data struggle to provide real-time predictions. This study introduces the Temporal Fusion Transformer (TFT) model to address these limitations by integrating sensor, location, and clinical data for predicting harmful behaviors. The TFT model’s advanced features, such as Variable Selection Networks and temporal attention mechanisms, make it particularly suitable for capturing complex time-series patterns and providing interpretable results in psychiatric settings.</div></div><div><h3>Methods</h3><div>Data from 145 patients across three hospitals were collected using wearable devices that tracked heart rate, movement, and location. The data were aggregated hourly, preprocessed to handle missing values, and standardized. A binary classification model using TFT was developed and evaluated with accuracy, recall, F1 score, and AUC. Bayesian optimization was employed for hyperparameter tuning, and 5-fold cross-validation was performed to ensure generalizability.</div></div><div><h3>Results</h3><div>The TFT model outperformed Multi-LSTM and Multi-GRU models, achieving 95.1% accuracy, 74.9% recall, an F1 score of 78.1, and an AUC of 0.863. The Variable Selection Network effectively identified key predictive factors, such as daily entropy and heart rate variability, improving interpretability. Incorporating location and biometric data enhanced prediction accuracy and enabled real-time risk assessments.</div></div><div><h3>Conclusion</h3><div>This study is the first to use the TFT model for predicting behavioral risks in psychiatric wards. The model’s ability to integrate diverse data sources, prioritize cirtical variables, and capture temporal dependencies make it highly suitable for psychiatric environments. While the TFT model performed well, challenges remain with recall due to the limited dataset. Future research will focus on expanding datasets, optimizing variable selection, and standardizing data through a multimodal Common Data Model (CDM) to further improve performance and clinical utility.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"198 ","pages":"Article 105870"},"PeriodicalIF":3.7,"publicationDate":"2025-03-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143642885","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Sensor-based prevention of falls and pressure ulcers: A scoping review","authors":"Anna Winkler ,&nbsp;Martin Pallauf ,&nbsp;Simon Krutter ,&nbsp;Patrick Kutschar ,&nbsp;Jürgen Osterbrink ,&nbsp;Nadja Nestler","doi":"10.1016/j.ijmedinf.2025.105878","DOIUrl":"10.1016/j.ijmedinf.2025.105878","url":null,"abstract":"<div><h3>Purpose</h3><div>Falls and pressure ulcers are serious complications impacting care quality in nursing homes. Sensor technologies can help prevent these adverse events through continuous monitoring and timely intervention. This scoping review, following JBI guidelines, evaluated the effects of sensor-based fall and pressure ulcer prevention in long-term care and the experiences of patients and healthcare professionals.</div></div><div><h3>Methods</h3><div>The review included primary studies, reviews, and protocols published from 2014 to 2023. Screening, data extraction, and quality appraisal were conducted independently by two authors using MMAT and JBI tools.</div></div><div><h3>Results</h3><div>A total of 31 studies were included: 22 on fall prevention, eight on pressure ulcer prevention, and one addressing both. User-based sensors were effective in preventing both falls and pressure ulcers. Accelerometers enhanced sensitivity for fall detection and adherence to repositioning protocols. Context-based sensors, such as Doppler, webcams, and Kinect, showed variable precision and false alarm rates, while range sensors demonstrated high precision. Context-based accelerometers were promising for pressure ulcer prevention, but pressure sensors provided inconsistent data. Additional manual assessments enhanced sensor data accuracy. Patients preferred non-obtrusive, user-friendly sensors, while healthcare professionals emphasized the need for seamless integration into care routines. Both groups valued real-time monitoring and alert capabilities, though privacy and data security remained concerns.</div></div><div><h3>Conclusions</h3><div>Sensor technologies show potential in enhancing patient safety and care quality in long-term care, though further refinement is needed for context-based sensors in pressure ulcer prevention. Integrating these technologies with standard care can improve outcomes, but addressing privacy and ethical issues is essential for broader acceptance.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"199 ","pages":"Article 105878"},"PeriodicalIF":3.7,"publicationDate":"2025-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143687184","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"LinkR: An open source, low-code and collaborative data science platform for healthcare data analysis and visualization","authors":"Boris Delange ,&nbsp;Benjamin Popoff ,&nbsp;Thibault Séité ,&nbsp;Antoine Lamer ,&nbsp;Adrien Parrot","doi":"10.1016/j.ijmedinf.2025.105876","DOIUrl":"10.1016/j.ijmedinf.2025.105876","url":null,"abstract":"<div><h3>Background</h3><div>The development of Clinical Data Warehouses (CDWs) has greatly increased access to big data in medical research. However, the lack of standardization among different data models hampers interoperability and, consequently, the research potential of these vast data resources. Moreover, data manipulation and analysis require advanced programming skills, a skill set that healthcare professionals often lack.</div></div><div><h3>Methods</h3><div>To address these issues, we created an open source, low-code and collaborative data science platform for manipulating, visualizing and analyzing healthcare data using graphical tools alongside an advanced programming interface. The software is based on the OMOP Common Data Model.</div></div><div><h3>Results</h3><div>LinkR enables users to generate studies using data imported from multiple sources. The software organizes the studies into two main sections: individual and population data sections. In the <em>individual data section</em>, user-friendly graphical tools allow users to customize data presentation, recreating the equivalent of a medical record, according to the needs of their study. The <em>population data section</em> is designed for conducting statistical analyses through both graphical and programming interfaces. The application also integrates a Git module, streamlining collaboration and facilitating shared data analysis across research centers. The platform was tested with datasets including the OMOP database (46,520 patients and over 36 million rows in the measurement table) during the InterHop datathon with 12 concurrent users. Usability testing yielded a median System Usability Scale (SUS) score of 75 [63.8–85.6], indicating high user satisfaction.</div></div><div><h3>Conclusion</h3><div>LinkR is a low-code data science platform that democratizes access, manipulation, and analysis of data from clinical data warehouses and facilitates collaborative work on healthcare data, using an open science approach.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"199 ","pages":"Article 105876"},"PeriodicalIF":3.7,"publicationDate":"2025-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143687195","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Impact of patients’ personality traits on digital health Adoption Strategies for family practices","authors":"Julian Beerbaum ,&nbsp;Sibylle Robens ,&nbsp;Leonard Fehring ,&nbsp;Achim Mortsiefer ,&nbsp;Sven Meister","doi":"10.1016/j.ijmedinf.2025.105880","DOIUrl":"10.1016/j.ijmedinf.2025.105880","url":null,"abstract":"<div><h3>Background</h3><div>Various governments highlight the relevance of digitalization in family practices; however, still some adoption barriers persist due to an inadequate understanding of why patients engage in digital use cases. Different studies show that personality traits influence how individuals assess digital use cases. Nevertheless, the effect of personality has not yet been tested in a family practice setting, even though family doctors are in an optimal position to use these personality insights via an empathetic communication approach in their direct patient interaction.</div></div><div><h3>Objective</h3><div>This paper aims to assess the impact of different personality traits on patients’ technology acceptance and derive implications for Digital Health Adoption Strategies of family practices – hence, what family doctors need to consider when influencing a patient’s decision to adopt a particular digital use case.</div></div><div><h3>Methods</h3><div>After reviewing the literature regarding the impact of personality on technology acceptance, we combined two established UTAUT and Big-Five questionnaires in a web-based survey. Recruiting a large cross-sectional sample of adults living in Germany, we conducted regression analyses to determine the effect of personality and sociodemographics on technology acceptance of four digital use cases in family practices.</div></div><div><h3>Results</h3><div>Our sample of 1,880 participants indicated that sociodemographics explained technology acceptance better than personality traits. Specifically, digital literacy, age and frequency of doctor visits affected people’s perception of different digital use cases while extraversion appeared as key personality trait in technology acceptance.</div></div><div><h3>Conclusion</h3><div>Family practices only need to consider personality traits selectively in developing Digital Health Adoption Strategies. Nevertheless, we argue that different patient personality profiles can guide family doctors in tailoring their communication while implementing digital use cases.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"198 ","pages":"Article 105880"},"PeriodicalIF":3.7,"publicationDate":"2025-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143631758","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 validation of an interpretable machine learning model for predicting in-hospital mortality for ischemic stroke patients in ICU","authors":"Xiao Luo ,&nbsp;Binghan Li ,&nbsp;Ronghui Zhu ,&nbsp;Yaoyong Tai ,&nbsp;Zongyu Wang ,&nbsp;Qian He ,&nbsp;Yanfang Zhao ,&nbsp;Xiaoying Bi ,&nbsp;Cheng Wu","doi":"10.1016/j.ijmedinf.2025.105874","DOIUrl":"10.1016/j.ijmedinf.2025.105874","url":null,"abstract":"<div><h3>Background</h3><div>Timely and accurate outcome prediction is essential for clinical decision-making for ischemic stroke patients in the intensive care unit (ICU). However, the interpretation and translation of predictive models into clinical applications are equally crucial. This study aims to develop an interpretable machine learning (IML) model that effectively predicts in-hospital mortality for ischemic stroke patients.</div></div><div><h3>Methods</h3><div>In this study, an IML model was developed and validated using multicenter cohorts of 3225 ischemic stroke patients admitted to the ICU. Nine machine learning (ML) models, including logistic regression (LR), K-nearest neighbors (KNN), naive Bayes (NB), decision tree (DT), support vector machine (SVM), random forest (RF), XGBoost, LightGBM, and artificial neural network (ANN), were developed to predict in-hospital mortality using data from the MIMIC-IV and externally validated in Shanghai Changhai Hospital. Feature selection was conducted using three algorithms. Model’s performance was assessed using area under the receiver operating characteristic (AUROC), accuracy, sensitivity, specificity and F1 score. Calibration curve and Brier score were used to evaluate the degree of calibration of the model, and decision curve analysis were generated to assess the net clinical benefit. Additionally, the SHapley Additive exPlanations (SHAP) method was employed to evaluate the risk of in-hospital mortality among ischemic stroke patients admitted to the ICU.</div></div><div><h3>Results</h3><div>Mechanical ventilation, age, statins, white blood cell, blood urea nitrogen, hematocrit, warfarin, bicarbonate and systolic blood pressure were selected as the nine most influential variables. The RF model demonstrated the most robust predictive performance, achieving AUROC values of 0.908 and 0.858 in the testing set and external validation set, respectively. Calibration curves also revealed a high consistency between observations and predictions. Decision curve analysis showed that the model had the greatest net benefit rate when the prediction probability threshold is 0.10 ∼ 0.80. SHAP was employed to interpret the RF model. In addition, we have developed an online prediction calculator for ischemic stroke patients.</div></div><div><h3>Conclusion</h3><div>This study develops a machine learning–based calculator to predict the probability of in-hospital mortality among patients with ischemic stroke in ICU. The calculator has the potential to guide clinical decision-making and improve the care of patients with ischemic stroke by identifying patients at a higher risk of in-hospital mortality.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"198 ","pages":"Article 105874"},"PeriodicalIF":3.7,"publicationDate":"2025-03-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143593311","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A qualitative study exploring electronic health record optimisation activities in English hospitals","authors":"Kathrin Cresswell ,&nbsp;Susan Hinder ,&nbsp;Robin Williams","doi":"10.1016/j.ijmedinf.2025.105868","DOIUrl":"10.1016/j.ijmedinf.2025.105868","url":null,"abstract":"<div><h3>Background</h3><div>Hospitals increasingly implement complex electronic health record (EHR) systems to improve quality, safety and efficiency. Whilst many aspects surrounding implementation and adoption processes have been researched, the benefits of such enterprise-wide systems may take decades to materialise. Existing work on optimisation processes has focused on technological, workflow and organisational aspects of optimisation within individual clinical settings, mostly in the United States of America. We here sought to explore how a range of hospitals with different EHR systems have approached the optimisation of EHRs over time and in relation to technology, socio-organisational and health system factors.</div></div><div><h3>Methods</h3><div>We conducted an in-depth qualitative interview study with technology leads from purposefully sampled hospitals across the country who had implemented a range of EHRs. We explored reflections on the journey of implementing and optimising systems over time, optimisation activities, and perceived lessons learned. Data were transcribed and analysed with NVivo 14 software, using the Technology, People, Organizations, and Macroenvironmental factors (TPOM) framework<!--> <!-->to facilitate coding.</div></div><div><h3>Results</h3><div>We interviewed 28 individuals from 21 sites with eight different types of EHRs. We observed various optimisation activities across different technological, social, organisational and health system factors. These included improving usability and information technology infrastructures; process optimisation of clinical and administrative workflows; organisational optimisation strategies and relationships with suppliers; and wider system factors such as the need for overall strategic direction and allocation of associated funding. Optimisation activities within these areas stood in some instances in contrast to one another. For example, national activities inhibited local optimisation efforts and organisational optimisation in some instances impacted adversely on usability.</div></div><div><h3>Conclusions</h3><div>This work emphasises that EHRs are not finished solutions but components of broader information systems needing continuous technological and organisational development. Effective optimisation requires a delicate balance between navigating technological affordances and characteristics to improve usability and organisational processes, as well as regional and national integration to achieve larger-scale interoperability.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"198 ","pages":"Article 105868"},"PeriodicalIF":3.7,"publicationDate":"2025-03-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143593422","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Patient consent for the secondary use of health data in artificial intelligence (AI) models: A scoping review","authors":"Khadijeh Moulaei ,&nbsp;Saeed Akhlaghpour ,&nbsp;Farhad Fatehi","doi":"10.1016/j.ijmedinf.2025.105872","DOIUrl":"10.1016/j.ijmedinf.2025.105872","url":null,"abstract":"<div><h3>Background</h3><div>The secondary use of health data for training Artificial Intelligence (AI) models holds immense potential for advancing medical research and healthcare delivery. However, ensuring patient consent for such utilization is paramount to uphold ethical standards and data privacy. Patient informed consent means patients are fully informed about how their data will be collected, used, and protected, and they voluntarily agree to allow their data to be used for AI models. In addition to formal consent frameworks, establishing a social license is critical to foster public trust and societal acceptance for the secondary use of health data in AI systems. This study examines patient consent practices in this domain.</div></div><div><h3>Method</h3><div>In this scoping review, we searched Web of Science, PubMed, and Scopus. We included studies in English that addressed the core issues of interest, namely, privacy, security, legal, and ethical issues related to the secondary use of health data in AI models. Articles not addressing the core issues, as well as systematic reviews, <em>meta</em>-analyses, books, letters, conference abstracts, and study protocols were excluded. Two authors independently screened titles, abstracts, and full texts, resolving disagreements with a third author. Data was extracted using a data extraction form.</div></div><div><h3>Results</h3><div>After screening 774 articles, a total of 38 articles were ultimately included in the review. Across these studies, a total of 178 barriers and 193 facilitators were identified. We consolidated similar codes and extracted 65 barriers and 101 facilitators, which we then categorized into four themes: “Structure,” “People,” “Physical system,” and “Task.” We identified notable emphasis on “Legal and Ethical Challenges” and “Interoperability and Data Governance.” Key barriers included concerns over privacy and security breaches, inadequacies in informed consent processes, and unauthorized data sharing. Critical facilitators included enhancing patient consent procedures, improving data privacy through anonymization, and promoting ethical standards for data usage.</div></div><div><h3>Conclusion</h3><div>Our study underscores the complexity of patient consent for the secondary use of health data in AI models, highlighting significant barriers and facilitators within legal, ethical, and technological domains. We recommend the development of specific guidelines and actionable strategies for policymakers, practitioners, and researchers to improve informed consent, ensuring privacy, trust, and ethical use of data, thereby facilitating the responsible advancement of AI in healthcare.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"198 ","pages":"Article 105872"},"PeriodicalIF":3.7,"publicationDate":"2025-03-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143642884","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 explainability for survival outcome in head and neck squamous cell carcinoma","authors":"Rasheed Omobolaji Alabi ,&nbsp;Antti A. Mäkitie ,&nbsp;Mohammed Elmusrati ,&nbsp;Alhadi Almangush ,&nbsp;Ylva Tiblom Ehrsson ,&nbsp;Göran Laurell","doi":"10.1016/j.ijmedinf.2025.105873","DOIUrl":"10.1016/j.ijmedinf.2025.105873","url":null,"abstract":"<div><h3>Background</h3><div>Diagnosis and treatment of head and neck squamous cell carcinoma (HNSCC) induces psychological variables and treatment-related toxicity in patients. The evaluation of outcomes is warranted for effective treatment planning and improved disease management. <strong>Objectives</strong>: This study aimed to build a prognostic system by combining clinicopathological parameters, treatment-related factors, and sociodemographic factors as integrative inputs to build a machine learning (ML) model to estimate the overall survival (OS) of patients with HNSCC. Furthermore, we explored the complementary prognostic potentials of these input parameters. We provide explainability and interpretability using Local Interpretable Model-agnostic Explanations (LIME) and SHapley Additive exPlanations (SHAP) techniques. <strong>Methods:</strong> A total of 419 patients with HNSCC were recruited from three University Hospitals in Sweden. We compared the performance of TabNet, a state-of-the-art deep learning algorithm for tabular data, with extreme gradient boosting (XGBoost) and voting ensemble to predict OS in patients with HNSCC. <strong>Results:</strong> Both TabNet and XGBoost showed comparable performance accuracies, with TabNet and XGBoost showing a performance accuracy of 88.1% each and voting ensemble showing an accuracy of 88.7%. The aggregate feature importance showed that p16 (a tumor suppressor protein that plays a crucial role in cell cycle regulation), cancer stage, hemoglobin, age at diagnosis, T class, N class, smoking pack-years, body mass index (BMI), treatment modality, erythrocyte count, and human papillomavirus (HPV) status were the most important parameters for the predictive ability of the model for OS. Furthermore, we found survival trends in this cohort by individually considering parameters such as p16, cancer stage, hemoglobin, age at diagnosis, HPV status, Tumor Nodal Metastasis staging, and socioeconomic factors (marital status, housing, and level of education). In addition, both the LIME and SHAP techniques showed the contribution of each feature to the prediction made by the model. <strong>Conclusions:</strong> The clinical implementation of an ML model can lead to individualized risk-based therapeutic decision-making. Therefore, validating these models with multi-institutional datasets and testing them in the context of clinical trials is warranted for safe clinical implementation.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"199 ","pages":"Article 105873"},"PeriodicalIF":3.7,"publicationDate":"2025-03-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143687192","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}