International Journal of Medical Informatics最新文献

{"title":"EPMA-related contributory factors to medication errors: development of a taxonomy to inform the optimisation strategy for an electronic patient record","authors":"Eileen C. Relihan ,&nbsp;Sinead M. Kelly","doi":"10.1016/j.ijmedinf.2025.105990","DOIUrl":"10.1016/j.ijmedinf.2025.105990","url":null,"abstract":"<div><h3>Background</h3><div>Optimisation of a hospital’s electronic prescribing and medication administration (EPMA) system requires a process of iterative refinement based on real world user experience. One source of this information is the medication error dataset generated from internal reporting systems. A process for analysis of this data is needed however, in order to derive meaningful insights over time.</div></div><div><h3>Objective</h3><div>To develop a taxonomy to facilitate the contemporaneous tracking and analysis of EPMA-related contributory factors to medication errors.</div></div><div><h3>Materials and Methods</h3><div>Medication errors reported over a 6-year period following implementation of a hospital EPMA system were analysed in order to identify incidents where the electronic system had been a contributory factor. A taxonomy was designed which captured the common themes and was suitable for incorporation into the reporting system.</div></div><div><h3>Conclusion</h3><div>A tool was designed for real-time recording and categorisation of EPMA-related contributory factors to medication errors. Data generated from the taxonomy is used to inform the optimisation strategy for the EPMA system.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"203 ","pages":"Article 105990"},"PeriodicalIF":3.7,"publicationDate":"2025-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144167898","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":"Lifting the wellbeing of adolescents and young adults with type 1 diabetes: A feasibility study of the LIFT app","authors":"Anna Serlachius ,&nbsp;Joanna McClintock ,&nbsp;Anna Boggiss ,&nbsp;Katie Babbott ,&nbsp;Chloe Hudson ,&nbsp;Paul Hofman ,&nbsp;Hiran Thabrew ,&nbsp;Alana Cavadino ,&nbsp;Craig Jefferies ,&nbsp;Ryan Paul ,&nbsp;Martin de Bock ,&nbsp;Jennifer Sherr","doi":"10.1016/j.ijmedinf.2025.105992","DOIUrl":"10.1016/j.ijmedinf.2025.105992","url":null,"abstract":"<div><h3>Background</h3><div>Adolescents and young adults (AYAs) with type 1 diabetes (T1D) have an increased risk of psychological distress. To address this, psychological support provided asynchronously via an app may be feasible. Our study aimed to explore feasibility and safety of the LIFT wellbeing app.</div></div><div><h3>Methods</h3><div>A 12-week single arm feasibility study was conducted with 59 AYAs aged 16–25 years recruited from New Zealand and the United States, and twenty-two support people (e.g. caregivers). Feasibility outcomes included retention, data completeness, user engagement and safety. Psychosocial and clinical outcomes (HbA1c and time-in-range) were assessed at baseline and 12 weeks.</div></div><div><h3>Results</h3><div>Retention and completion of self-reported outcome measures was &gt; 80 % for AYAs and support people. Users reported good engagement with the app. No adverse events occurred. Psychosocial outcome measures showed promising changes from baseline to 12-weeks.</div></div><div><h3>Conclusions</h3><div>LIFT was deemed engaging, safe and feasible with promising preliminary changes in psychological outcomes.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"203 ","pages":"Article 105992"},"PeriodicalIF":3.7,"publicationDate":"2025-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144167899","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":"Utilizing artificial intelligence and medical experts to identify predictors for common diagnoses in dyspneic adults: A cross-sectional study of consecutive emergency department patients from Southern Sweden","authors":"Ellen T. Heyman ,&nbsp;Awais Ashfaq ,&nbsp;Ulf Ekelund ,&nbsp;Mattias Ohlsson ,&nbsp;Jonas Björk ,&nbsp;Alexander Marcel Schubert ,&nbsp;Markus Lingman ,&nbsp;Ardavan M. Khoshnood","doi":"10.1016/j.ijmedinf.2025.105969","DOIUrl":"10.1016/j.ijmedinf.2025.105969","url":null,"abstract":"<div><h3>Objective</h3><div>Half of all adult emergency department (ED) visits with a complaint of dyspnea involve acute heart failure (AHF), exacerbation of chronic obstructive pulmonary disease (eCOPD), or pneumonia, which are often misdiagnosed. We aimed to create an artificial intelligence (AI) diagnostic decision support tool to detect patients with AHF, eCOPD, and pneumonia among dyspneic adults at the beginning of their ED visit.</div></div><div><h3>Methods</h3><div>In this cross-sectional study, we included all ED visits of patients 18 years or older with dyspnea at two regional Swedish EDs 07/01/2017–12/31/2019. In-hospital or ED discharge notes were used as outcome labels, with a subset manually reviewed by experts. We analyzed data from a complete regional healthcare system, along with socioeconomic factors, using Hierarchical Attention Networks. Each patient displayed a unique set of variables important for diagnosing dyspnea. All patients’ unique variable sets were aggregated into a variable list. The top 100, 50, and 20 variables were tested in a simpler CatBoost model. Finally, performance was compared after adding medical expertise to the AI model.</div></div><div><h3>Results</h3><div>We included 10,869 visits, with 15.1% having AHF, 13.6% eCOPD, and 13.1% pneumonia. The median number of variables per unique ED visit was 187 (IQR 111–307). Aggregating the unique sets of variables resulted in a cohort list of 2,064 variables. The median micro AUROC was 87.8% (2.5–97.5 percentile; 86.4–89.3%). Age, ECGs, previous diagnoses, and medication were considered important by the AI model, while sex, vital signs, and socioeconomic factors were deemed almost non-predictive. Using the top 20 AI-selected variables, the AUROC was 86.6% (85.1–88.1%). Adding human medical expertise did not significantly change the AUROC.</div></div><div><h3>Conclusion</h3><div>Based on the analysis of a high-dimensional dataset, we designed a lightweight 20-variable machine learning model that can early and effectively diagnose AHF, eCOPD, and pneumonia among ED patients with dyspnea.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"202 ","pages":"Article 105969"},"PeriodicalIF":3.7,"publicationDate":"2025-05-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144170167","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":"The clinical effectiveness in telemedicine in Trauma: A systematic review and meta-analysis","authors":"Thamer Nouh ,&nbsp;Mishary Shalhoub ,&nbsp;Ahmed Alburakan ,&nbsp;Nawaf Alshahwan ,&nbsp;Suha Kaaki ,&nbsp;Mohannad Hemdi ,&nbsp;Lama Al zelfawi ,&nbsp;Ebtesam Almajed ,&nbsp;Ghadah Aldayel ,&nbsp;Alanood Alharthi ,&nbsp;Reem AlSarhan ,&nbsp;Reem Altamimi ,&nbsp;Kayan Alotaibi ,&nbsp;Saleh Aldeligan ,&nbsp;Abdulrahman Alsughayyir ,&nbsp;Ghada Alharbi ,&nbsp;Hassan Mashbari","doi":"10.1016/j.ijmedinf.2025.105986","DOIUrl":"10.1016/j.ijmedinf.2025.105986","url":null,"abstract":"<div><h3>Background</h3><div>Telemedicine (TM) has emerged as a transformative tool in trauma care, enabling remote consultation and guidance for healthcare providers, especially in resource-limited settings. This review evaluates TM’s effectiveness in trauma care by assessing its impact on clinical outcomes, mortality, injury severity, and healthcare resource utilization.</div></div><div><h3>Patients and Methods</h3><div>A systematic literature search was performed across PubMed, Google Scholar, Science Direct, ProQuest, Medline, and Web of Science and included randomized controlled trials and cohort studies involving adult trauma patients. Extracted data encompassed demographics, injury characteristics, and clinical outcomes.</div></div><div><h3>Results</h3><div>Analysis of 25 studies (n = 45,097 patients) revealed that TM was associated with higher Injury Severity Scores (ISS), likely due to improved detection and documentation of severe injuries through specialist input. Increased blood transfusion rates were observed, which may reflect more accurate triage and timely intervention facilitated by TM, though confounding factors (e.g., case mix) cannot be ruled out. Both patient and physician satisfaction rates were high. However, TM demonstrated no statistically significant effect on mortality (p &gt; 0.05), transfer times, or emergency department (ED) utilization—a finding consistent with some prior studies but warranting further investigation, given the heterogeneity in protocols.</div></div><div><h3>Conclusion</h3><div>TM enhances trauma care coordination in rural and underserved areas by improving access to specialist expertise, which may explain observed benefits like higher ISS detection and transfusion rates. However, its inconsistent impact on mortality and resource use underscores the need for standardized implementation protocols. Future integration should optimize TM workflows (e.g., reducing decision delays) and pair them with targeted resource allocation to maximize outcomes.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"202 ","pages":"Article 105986"},"PeriodicalIF":3.7,"publicationDate":"2025-05-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144116862","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":"Performance of multimodal prediction models for intracerebral hemorrhage outcomes using real-world data","authors":"Koutarou Matsumoto ,&nbsp;Masahiro Suzuki ,&nbsp;Kazuaki Ishihara ,&nbsp;Koki Tokunaga ,&nbsp;Katsuhiko Matsuda ,&nbsp;Jenhui Chen ,&nbsp;Shigeo Yamashiro ,&nbsp;Hidehisa Soejima ,&nbsp;Naoki Nakashima ,&nbsp;Masahiro Kamouchi","doi":"10.1016/j.ijmedinf.2025.105989","DOIUrl":"10.1016/j.ijmedinf.2025.105989","url":null,"abstract":"<div><h3>Background</h3><div>We aimed to develop and validate multimodal models integrating computed tomography (CT) images, text and tabular clinical data to predict poor functional outcomes and in-hospital mortality in patients with intracerebral hemorrhage (ICH). These models were designed to assist non-specialists in emergency settings with limited access to stroke specialists.</div></div><div><h3>Methods</h3><div>A retrospective analysis of 527 patients with ICH admitted to a Japanese tertiary hospital between April 2019 and February 2022 was conducted. Deep learning techniques were used to extract features from three-dimensional CT images and unstructured data, which were then combined with tabular data to develop an L1-regularized logistic regression model to predict poor functional outcomes (modified Rankin scale score 3–6) and in-hospital mortality. The model’s performance was evaluated by assessing discrimination metrics, calibration plots, and decision curve analysis (DCA) using temporal validation data.</div></div><div><h3>Results</h3><div>The multimodal model utilizing both imaging and text data, such as medical interviews, exhibited the highest performance in predicting poor functional outcomes. In contrast, the model that combined imaging with tabular data, including physiological and laboratory results, demonstrated the best predictive performance for in-hospital mortality. These models exhibited high discriminative performance, with areas under the receiver operating curve (AUROCs) of 0.86 (95% CI: 0.79–0.92) and 0.91 (95% CI: 0.84–0.96) for poor functional outcomes and in-hospital mortality, respectively. Calibration was satisfactory for predicting poor functional outcomes, but requires refinement for mortality prediction. The models performed similar to or better than conventional risk scores, and DCA curves supported their clinical utility.</div></div><div><h3>Conclusion</h3><div>Multimodal prediction models have the potential to aid non-specialists in making informed decisions regarding ICH cases in emergency departments as part of clinical decision support systems. Enhancing real-world data infrastructure and improving model calibration are essential for successful implementation in clinical practice.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"202 ","pages":"Article 105989"},"PeriodicalIF":3.7,"publicationDate":"2025-05-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144116790","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":"Measuring information technology expenditures of hospitals for cross-institutional comparisons: A scoping review","authors":"Luka Leon Reincke,&nbsp;Jan-David Liebe","doi":"10.1016/j.ijmedinf.2025.105968","DOIUrl":"10.1016/j.ijmedinf.2025.105968","url":null,"abstract":"<div><h3>Background</h3><div>Information technology (IT) expenditure is a structural prerequisite for digital transformation in hospitals. Despite increasing research into the level of digital maturity and increased political initiatives to finance hospital digitalization, this financial dimension remains methodologically unexplored. It can be assumed that one reason for this is the inconsistency of definitions, fragmented data and a lack of standardized reporting, which hinders cross-institutional comparisons. This makes it difficult to evaluate investments in digitalization, especially in the context of national funding programs. At the same time, the scientific investigation of the effects of IT expenditure on the success of digital transformation in comparison to other influencing factors is difficult to answer.</div></div><div><h3>Objectives</h3><div>This review takes a methodological perspective and aims to (I) provide an overview of how hospital IT expenditures of hospitals are measured in the context of cross-institutional comparisons, (II) identify methodological challenges and (III) propose a conceptual framework that enables cross-institutional comparisons and thus helps future evaluation and research projects.</div></div><div><h3>Methods</h3><div>This scoping review followed Arksey and O'Malley's framework and adhered to PRISMA-ScR guidelines. Relevant studies were systematically identified through PubMed, Web of Science and Scopus. Analysis was based on a standardized scheme capturing data sources, IT-related metrics, control variables and methodological challenges.</div></div><div><h3>Results</h3><div>Ten studies were identified that measure IT expenditures in hospitals for cross-institutional comparison. Three main metrics were used, IT capital (n=9), IT personnel (n=6) and IT operating expenses (n=4), but with inconsistent definitions. Most studies relied on secondary data, often outdated and difficult to access, while primary or mixed-methods approaches were rare. Structural control variables were commonly applied; macroeconomic and regulatory factors were seldom considered. A conceptual framework was developed to address three main challenges: limited data availability, heterogeneous metrics and inconsistent use of contextual variables.</div></div><div><h3>Conclusion</h3><div>This scoping review highlights considerable methodological heterogeneity in the cross-institutional measurement of IT expenditures of hospitals, particularly concerning data sources, expenditure metrics and control variables. The findings emphasize the need for theory-informed, empirically feasible and context-sensitive approaches to ensure valid cross-institutional comparisons. The conceptual framework proposed in this review provides methodological guidance for future research and contributes to the standardization of IT expenditure measurement in the context of cross-institutional comparisons.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"202 ","pages":"Article 105968"},"PeriodicalIF":3.7,"publicationDate":"2025-05-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144116861","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":"Validating the information technology (IT) implementation framework to Implement mHealth technology for consumers: A case study of the Sense2Quit app for smoking cessation","authors":"Maeve Brin ,&nbsp;Sydney Fontalvo ,&nbsp;David Hu ,&nbsp;Patricia Cioe ,&nbsp;Ming-Chun Huang ,&nbsp;Wenyao Xu ,&nbsp;Rebecca Schnall","doi":"10.1016/j.ijmedinf.2025.105977","DOIUrl":"10.1016/j.ijmedinf.2025.105977","url":null,"abstract":"<div><h3>Objective</h3><div>The goal of this paper was to understand the applicability of the Information Technology (IT) Implementation Framework, a multi-level approach to identify factors that impede or promote IT usage, for incorporating a mHealth technology for consumers in the community setting.</div></div><div><h3>Methods</h3><div>A case study of the implementation of the Sense2Quit App for smoking cessation among people living with HIV was examined to parse out the factors within the framework that are applicable to mHealth technology and the factors that may need modification for use of this framework within this context.</div></div><div><h3>Results</h3><div>Findings suggest that phases two through five of the framework were applicable to our study and phase one was not.</div></div><div><h3>Conclusion</h3><div>Findings support the use of the theory for implementation of mHealth technology for promoting consumer health at the community level. This use case may be useful for stakeholders evaluating implementation of mHealth for patients with chronic conditions as it highlights the need to identify preferences of app specifications, personal habits, and various factors such as confidentiality and digital literacy which may challenge sustained usage.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"202 ","pages":"Article 105977"},"PeriodicalIF":3.7,"publicationDate":"2025-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144123516","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":"Digital applications to support self-management of multimorbidity: A scoping review","authors":"Lucy Smith,&nbsp;Glenn Simpson,&nbsp;Sian Holt,&nbsp;Hajira Dambha-Miller","doi":"10.1016/j.ijmedinf.2025.105988","DOIUrl":"10.1016/j.ijmedinf.2025.105988","url":null,"abstract":"<div><h3>Introduction</h3><div>Multimorbidity, defined as the co-occurrence of two or more long-term conditions, is increasing rapidly and poses challenges for healthcare systems. Advances in digital technologies offer solutions by facilitating personalised, scalable care interventions that empower individuals to manage their conditions more effectively. These applications have potential to improve access to care, enhance patient engagement, and support tailored approaches to self-management.</div></div><div><h3>Objectives</h3><div>This scoping review aims to synthesise current evidence on the use of digital applications for self-management in adults with multimorbidity.</div></div><div><h3>Methods</h3><div>A scoping review was conducted, systematically searching PubMed, Web of Science, OVID, CINAHL, EMBASE, and additional manual searches. Boolean operators and targeted key terms were employed to retrieve relevant studies from database inception to 16th January 2024.</div></div><div><h3>Results</h3><div>The search yielded 1,974 articles, of which 31 met the inclusion criteria. Digital applications for self-management in multimorbidity demonstrated high acceptability and varying efficacy. Key benefits included improved communication, symptom tracking, and autonomy. Barriers included privacy concerns, additional patient burden, and engagement challenges. Socio-demographics, self-efficacy, and digital literacy influenced both barriers and facilitators to tool usage. Theoretical models underpinning digital applications were limited. Older adults and the working-age population were rarely included.</div></div><div><h3>Conclusion</h3><div>The current evidence base does not fully address the needs of older adults with low digital literacy or working-age populations with multimorbidity. Our model highlights the importance of broader contextual mechanisms in digital tool adoption. Future research should prioritise theory-driven tool development tailored to disease clusters and aligned with sociodemographic profiles, health risks, and social care needs. Addressing these gaps could improve self-management and health outcomes for high-risk populations.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"202 ","pages":"Article 105988"},"PeriodicalIF":3.7,"publicationDate":"2025-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144135029","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":"Artificial intelligence in tobacco control: A systematic scoping review of applications, challenges, and ethical implications","authors":"David B. Olawade ,&nbsp;Charity A. Aienobe-Asekharen","doi":"10.1016/j.ijmedinf.2025.105987","DOIUrl":"10.1016/j.ijmedinf.2025.105987","url":null,"abstract":"<div><h3>Background</h3><div>Tobacco use remains a significant global health challenge, contributing substantially to preventable morbidity and mortality. Despite established interventions, outcomes vary due to scalability issues, resource constraints, and limited reach.</div></div><div><h3>Objective</h3><div>To systematically explore current artificial intelligence (AI) applications within tobacco control, highlighting their usefulness, benefits, limitations, and ethical implications.</div></div><div><h3>Method</h3><div>This scoping review followed the Arksey and O’Malley framework and PRISMA-ScR guidelines. Five major databases (PubMed, Scopus, Web of Science, IEEE Xplore, and PsycINFO) were searched for articles published between January 2010 and March 2025. From 1,172 initial records, 57 studies met inclusion criteria after screening.</div></div><div><h3>Results</h3><div>AI-driven tools, including machine learning and natural language processing, effectively monitor social media for emerging tobacco trends and personalize smoking cessation interventions. Applications were predominantly focused on predictive modelling (using algorithms like XGBoost and SVM to predict e-cigarette use and relapse risk), cessation support (employing chatbots and reinforcement learning to improve accessibility), and social media surveillance (detecting synthetic nicotine promotions and analysing vaping trends). Approximately 22% of studies aligned with WHO FCTC Article 13 (tobacco advertising regulation), while 45% supported Article 14 (cessation services). However, tobacco industry interference remains a critical challenge, with AI technologies exploited to undermine public health initiatives, target vulnerable populations, and manipulate policy discussions. Critical issues including algorithmic bias, privacy concerns, interpretability challenges, and data quality must be addressed to ensure positive impact.</div></div><div><h3>Conclusion</h3><div>AI holds considerable promise for extending tobacco control if implemented ethically, transparently, and collaboratively. Future directions emphasize explainable AI development, integration of real-time intervention systems, global data inclusion, and robust cross-sector collaboration. While the current landscape shows a laudable start, it reflects the need for more diverse skill sets to fully harness AI’s extensive prospects for tobacco control and achieving tobacco endgame goals.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"202 ","pages":"Article 105987"},"PeriodicalIF":3.7,"publicationDate":"2025-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144105132","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":"Show and tell: A critical review on robustness and uncertainty for a more responsible medical AI","authors":"Luca Marconi ,&nbsp;Federico Cabitza","doi":"10.1016/j.ijmedinf.2025.105970","DOIUrl":"10.1016/j.ijmedinf.2025.105970","url":null,"abstract":"<div><div>This critical review explores two interrelated trends: the rapid increase in studies on machine learning (ML) applications within health informatics and the growing concerns about the reproducibility of these applications across different healthcare settings. Addressing these concerns necessitates acknowledging the uncertainty inherent in evaluating medical decision support systems. Therefore, we emphasize the importance of external validation and robustness assessment of the underlying ML models to better estimate their performance across diverse real-world scenarios.</div><div>To raise awareness among health practitioners and ML researchers, we advocate for the widespread adoption of external validation practices and uncertainty quantification techniques. Our survey of specialized literature reveals that fewer than 4% of studies published in high-impact medical informatics journals over the past 13 years have validated their systems using data from settings different from those that provided the training data. This low percentage is incompatible with responsible research, given the potential risks posed by unreliable ML models in healthcare.</div><div>Raising the standards for medical AI evaluation is crucial to improving practitioners' understanding of the potential and limitations of decision support systems in real-world settings. It is essential that uncertainty is not hidden in studies aimed at advancing knowledge in this field.</div></div>","PeriodicalId":54950,"journal":{"name":"International Journal of Medical Informatics","volume":"202 ","pages":"Article 105970"},"PeriodicalIF":3.7,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144139031","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}