JMIR Cardio最新文献

{"title":"Machine Learning Models for Mortality Prediction in Intensive Care Unit Patients With Ischemic Stroke Associated With Intracranial Artery Stenosis: Retrospective Cohort Study.","authors":"Kun Zhang, Ruomeng Chen, Jingyi Yang, Yan Yan, Lijuan Liu, Chaoyue Meng, Peifang Li, Guoying Xing, Xiaoyun Liu","doi":"10.2196/82042","DOIUrl":"10.2196/82042","url":null,"abstract":"<p><strong>Background: </strong>Mortality prediction in intensive care unit (ICU) patients with ischemic stroke complicated by intracranial artery stenosis or occlusion remains difficult. Conventional scoring systems often lack discriminatory power and fail to provide individualized risk estimates. Machine learning approaches have been increasingly explored to integrate diverse clinical features for prognostic modeling.</p><p><strong>Objective: </strong>This study aims to develop and evaluate machine learning models for individualized mortality prediction in ICU patients with ischemic stroke associated with intracranial artery stenosis or occlusion.</p><p><strong>Methods: </strong>Using the Medical Information Mart for Intensive Care IV (MIMIC-IV) database, we conducted a retrospective cohort study including 5280 adult ICU patients identified through International Classification of Diseases, Ninth and Tenth Revision (ICD-9/10) codes. Mortality status was determined based on the presence of a recorded date of death (dod) in the MIMIC-IV database. Patients with a documented dod were classified as deceased, whereas those without a recorded dod were classified as nondeceased. The primary outcome was all-cause mortality as recorded in the MIMIC-IV database, defined by the presence of a documented dod. Patients were randomly split into training (n=3696, 70%) and testing (n=1584, 30%) cohorts. Missing value imputation, correlation reduction, and multistep supervised feature selection (gradient boosting, BorutaShap, recursive feature elimination with cross-validation, LassoCV, and chi-square analysis) were performed exclusively within the training set and subsequently applied to the test set, resulting in 35 retained predictive features. Eight machine learning models-including light gradient boosting machine (LightGBM), Bagging (bootstrap aggregating), random forest, logistic regression, support vector machine, gradient boosting, adaptive boosting, and k-nearest neighbors-were trained with hyperparameter optimization using RandomizedSearchCV. Model performance was evaluated using area under the curve, accuracy, recall, precision, F1-score, and calibration curves. Shapley additive explanations were used for global and individual-level interpretability.</p><p><strong>Results: </strong>LightGBM, Bagging, and logistic regression demonstrated comparable discrimination, achieving an area under the curve of approximately 0.82-0.83 and accuracy above 73% on the independent test set. LightGBM demonstrated balanced performance (recall 0.70; precision 0.72) and good calibration. Shapley additive explanations analysis identified acute physiology score III, suspected infection, Charlson comorbidity index, age, weight on admission, and red cell distribution width as the most influential predictors. Overall, higher physiological severity, greater comorbidity burden, and older age were consistently associated with increased observed mortality risk.</p><p><strong>Conclus","PeriodicalId":14706,"journal":{"name":"JMIR Cardio","volume":"10 ","pages":"e82042"},"PeriodicalIF":2.2,"publicationDate":"2026-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12931835/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147284008","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Impact of the Cardio-Meds Mobile App on Heart Failure Knowledge and Medication Adherence: Pilot Randomized Controlled Trial.","authors":"Victor Buswell, Emmanuelle Massie, Elena Tessitore, Lisa Simioni, Guillaume Guebey, Hamdi Hagberg, Aurélie Schneider-Paccot, Samaksha Pant, Katherine Blondon, Liliane Gschwind, Frederic Ehrler, Philippe Meyer","doi":"10.2196/83022","DOIUrl":"https://doi.org/10.2196/83022","url":null,"abstract":"<p><strong>Background: </strong>Heart failure (HF) is a prevalent chronic condition for which optimal management depends not only on guideline-directed medical therapy but also on patients' understanding of their disease, recognition of warning signs, and sustained medication adherence, which remains challenging in routine care. Mobile health interventions may support therapeutic education and self-management; however, many available apps lack validated content and local relevance. Cardio-Meds is a mobile app developed at Geneva University Hospitals to support HF self-management through structured educational content, interactive quizzes, medication lists with reminders, and tools for monitoring weight and vital signs.</p><p><strong>Objective: </strong>This study aims to evaluate the impact of a 30-day Cardio-Meds intervention on HF knowledge and medication adherence in patients with HF with reduced or mildly reduced ejection fraction.</p><p><strong>Methods: </strong>We conducted a single-center, pilot randomized controlled trial in patients followed at the outpatient HF clinic or enrolled in cardiac rehabilitation at Geneva University Hospitals in 2024. Eligible participants had HF with a left ventricular ejection fraction less than 50%, were receiving HF-specific pharmacotherapy, speak French, and owned a smartphone. Participants were recruited by phone and randomized to Cardio-Meds use for 30 days, a self-guided intervention with a single standardized technical support call. Outcomes were self-assessed using standardized questionnaires: HF knowledge and self-management using the Dutch Heart Failure Knowledge Scale (DHFKS; score range 0-15); medication adherence using the Basel Assessment of Adherence to Immunosuppressive Medication Scale, covering initiation, implementation, and persistence; and usability in the intervention group using the System Usability Scale (score range 0-100). Between-group differences in DHFKS scores were analyzed using analysis of covariance adjusted for baseline values.</p><p><strong>Results: </strong>A total of 49 participants were included (25 intervention, 24 control); 78% (n=38) were male, and the mean age was 62 (SD 11.4) years. In the intervention group, median app usage was 123 (IQR 74-273) minutes, with a median of 43 (IQR 19-85) logins. Mean baseline DHFKS scores were similar between groups (intervention 11.1, SD 2.4 vs control 10.5, SD 2.9). At 30 days, mean scores increased significantly in the intervention group (12.4, SD 2.4; mean change +1.3; P<.001) and remained stable in the control group (10.4, SD 3; mean change -0.1; P=.82), with a significant adjusted between-group difference of +1.3 points (P<.001). No significant between-group differences were observed for medication adherence. Usability was high, with a mean score of 84.3 (SD 15), and 64% (16/25) of intervention participants reported that they would continue using the app.</p><p><strong>Conclusions: </strong>In a stable ambulatory HF population,","PeriodicalId":14706,"journal":{"name":"JMIR Cardio","volume":"10 ","pages":"e83022"},"PeriodicalIF":2.2,"publicationDate":"2026-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12928692/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147276064","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Perceived Potential and Challenges of Supporting Coronary Artery Disease Treatment Decisions With AI: Qualitative Study.","authors":"Khara Sauro, Bishnu Bajgain, Cody van Rassel, Bryan Har, Robert Welsh, Joon Lee","doi":"10.2196/81303","DOIUrl":"10.2196/81303","url":null,"abstract":"<p><strong>Background: </strong>Coronary revascularization decision-making for patients with coronary artery disease (CAD) can be complex and challenging. Artificial intelligence (AI) has the potential to improve this decision-making by bringing data-driven insights to the point of care.</p><p><strong>Objective: </strong>We aimed to elicit, collect, and analyze various stakeholders' perceived potential and challenges related to developing, implementing, and adopting AI-based CAD treatment decision support systems.</p><p><strong>Methods: </strong>A facilitated small-group discussion method, known as a World Café, was conducted with general cardiologists, interventional cardiologists, cardiac surgeons, patients, caregivers, health system administrators, and industry representatives. One-on-one interviews were conducted for participants who could not attend the World Café. Perceived potential and challenges of AI-based CAD treatment decision support systems were solicited by asking participants three broad questions: (1) What is most challenging about revascularization decision-making? (2) How could an AI tool be integrated into the existing clinical workflow? (3) What are the critical components that need to be considered when developing the AI tool? Thematic analysis was performed to identify themes from the data.</p><p><strong>Results: </strong>Nine participants completed the World Café, and 3 participants completed the one-on-one interviews. Five main themes emerged: (1) evidence-based care, (2) workload and resources, (3) data requirements (subthemes: patient-centered approach, evidence-based AI, and data integration), (4) tool characteristics (subthemes: end user built; generation and presentation of decision support information; user-friendliness and accessibility; and system logic, reasoning, and data privacy), and (5) incorporation into clinical workflow (subthemes: AI as an opportunity to improve care and knowledge translation).</p><p><strong>Conclusions: </strong>While health care providers aim to provide evidence-based care, CAD treatment decision-making can often be subjective due to the limited applicability of clinical practice guidelines and randomized controlled trial evidence to individual patients. AI-based clinical decision support systems may be an effective solution if the development and implementation focus on the issues identified by end users in this study (patient preference, data privacy, integration with clinical information systems, transparency, and usability).</p>","PeriodicalId":14706,"journal":{"name":"JMIR Cardio","volume":"10 ","pages":"e81303"},"PeriodicalIF":2.2,"publicationDate":"2026-02-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12880591/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146131792","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Mindfulness-Based Self-Management Program Using a Mobile App for Patients With Pulmonary Hypertension: Single-Arm Feasibility Study.","authors":"Yuka Takita, Junko Morishita, Sunre Park, Ayumi Goda, Takumi Inami, Hanako Kikuchi, Takashi Kohno, Masaharu Kataoka, Daisuke Fujisawa","doi":"10.2196/79639","DOIUrl":"10.2196/79639","url":null,"abstract":"<p><strong>Background: </strong>Mindfulness-based interventions have been applied across various chronic illnesses, but no tailored program exists for individuals with pulmonary hypertension (PH).</p><p><strong>Objective: </strong>This study aimed to develop and evaluate the feasibility of a mindfulness-based self-management program for patients with PH, delivered online to accommodate their limited mobility.</p><p><strong>Methods: </strong>A single-arm pre-post study was conducted using an 8-session, weekly videoconference program incorporating PH self-management education and elements of mindfulness-based cognitive therapy. A mobile app linked to an Apple Watch was used to support symptom monitoring and mindfulness awareness. Outcomes included PH-related symptoms, quality of life (emPHasis-10), depression (Patient Health Questionnaire-9 [PHQ-9]), anxiety (Generalized Anxiety Disorder 7-item scale [GAD-7]), resilience (Connor-Davidson Resilience Scale [CD-RISC]), and loneliness (UCLA Loneliness Scale-short version). Assessments occurred at baseline, week 4, and program completion. Exit interviews explored perceived changes and experiences.</p><p><strong>Results: </strong>Twelve participants (mean age 41.8, SD 10.5 years; range 26-56 years) were enrolled, and 9 completed the program (75% retention). Participants valued the online format and Apple Watch integration, while noting a need for optional on-demand sessions. Qualitative analysis identified themes such as increased self-awareness, use of meditation for pain management, and enhanced self-compassion. Quantitative analysis showed significant changes across 3 time points (baseline, week 4, and week 8) for emPHasis-10 (χ²₂=9.74; P=.008) and CD-RISC (χ²₂=7.27; P=.03). Trends toward change were observed for PHQ-9 (χ²₂=4.75; P=.09) and GAD-7 (χ²₂=5.07; P=.08), but week 12 data were limited (n=5). No significant changes in loneliness were observed.</p><p><strong>Conclusions: </strong>The program appeared to support patients with PH in managing symptoms and emotions and suggested potential improvements in quality of life. These preliminary findings warrant evaluation in a future randomized controlled trial.</p>","PeriodicalId":14706,"journal":{"name":"JMIR Cardio","volume":"10 ","pages":"e79639"},"PeriodicalIF":2.2,"publicationDate":"2026-02-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12871575/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146119004","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Impact of the Cardio-Meds mobile app on heart failure knowledge and medication adherence: a pilot randomized controlled trial.","authors":"Victor Buswell, Emmanuelle Massie, Elena Tessitore, Lisa Simioni, Guillaume Guebey, Hamdi Hagberg, Aurélie Schneider-Paccot, Samaksha Pant, Katherine Blondon, Liliane Gschwind, Frederic Ehrler, Philippe Meyer","doi":"10.2196/83022","DOIUrl":"https://doi.org/10.2196/83022","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Heart failure (HF) is a prevalent chronic condition associated with substantial morbidity, mortality, and healthcare utilization. Optimal management depends not only on guideline-directed medical therapy but also on patients' understanding of their disease, recognition of warning signs, and sustained medication adherence, areas that remain challenging in routine care, particularly in polymorbid patients with complex treatment regimens. Mobile health interventions may provide scalable support for therapeutic education and self-management; however, many available applications lack validated content and local relevance. Cardio-Meds is a mobile application developed at Geneva University Hospitals to support HF self-management through structured educational content, interactive quizzes with feedback, medication lists with optional reminders and intake confirmation, and tools for monitoring weight and vital signs.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Objective: &lt;/strong&gt;To evaluate the impact of a 30-day Cardio-Meds intervention on HF knowledge and self-management, and on medication adherence, in patients with HF with reduced or mildly reduced ejection fraction.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;We conducted a single-centre, pilot randomized controlled trial in patients followed at the outpatient HF clinic or enrolled in cardiac rehabilitation at Geneva University Hospitals between March and November 2024. Eligible participants had HF with left ventricular ejection fraction &lt;50%, were receiving HF-specific pharmacotherapy, were able to communicate in French, and owned a smartphone. Participants were recruited by phone and were randomized to Cardio-Meds use for 30 days, a self-guided intervention with a single standardized technical support call, plus usual care or to usual care alone. The outcomes were self-assessed using standardized questionnaires. HF knowledge and self-management were assessed at baseline and 30 days using the Dutch Heart Failure Knowledge Scale (DHFKS; score range 0-15). Medication adherence was evaluated using the Basel Assessment of Adherence to Immunosuppressive Medication Scale (BAASIS®), covering initiation, implementation, and persistence. Usability in the intervention group was assessed using the System Usability Scale (SUS; score range 0-100). Between-group differences in DHFKS scores were analysed using analysis of covariance adjusted for baseline values.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;A total of 49 participants were included (25 intervention, 24 control; 78% male; mean age 62±11.4 years). In intervention group, median app usage was 123 minutes (IQR 74-273), with median of 43 logins (IQR 19-85). Baseline DHFKS scores were similar between groups (intervention 11.1±2.4 vs control 10.5±2.9). At 30 days, DHFKS scores increased significantly in the intervention group (12.4±2.4; mean change +1.3, p&lt;0.001) and remained stable in the control group (10.4±3.0; mean change -0.1, p=0.82), with a significant adjusted between-gr","PeriodicalId":14706,"journal":{"name":"JMIR Cardio","volume":" ","pages":""},"PeriodicalIF":2.2,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146029538","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluation of a community-based SMS support program for cardiovascular patients from 2020 - 2024: The HeartHealth program.","authors":"Brodie Sheahen, Liliana Laranjo, Ritu Trivedi, Tim Shaw, Gopal Sivagangabalan, James Chong, Aravinda Thiagalingam, Sarah Zaman, Pierre Qian, Anupama Balasuriya Indrawansa, Clara Kayei Chow","doi":"10.2196/68896","DOIUrl":"10.2196/68896","url":null,"abstract":"<p><strong>Background: </strong>The HeartHealth program is a six-month SMS message-based support program offered to patients with a recent cardiovascular hospitalisation or recent cardiovascular clinic visit in Western Sydney, Australia. Its customised content focuses on cardiovascular risk factors, lifestyle, treatments and general heart health information.</p><p><strong>Objective: </strong>To evaluate the implementation of the HeartHealth program.</p><p><strong>Methods: </strong>A mixed-methods study was conducted assessing program reach, effectiveness, implementation and maintenance using program data, participant feedback surveys and staff focus-group discussions. Consecutive adult patients who had attended cardiology clinics or had been discharged from cardiology hospitalisation at Westmead Hospital, between April 2020 and April 2024, were included in the analysis. Content analysis was utilised to interpret the qualitative data.</p><p><strong>Results: </strong>A total of 23095 patients were invited, 8804 (38.1%; 8804/23095) enrolled into the program, and 7964 (90.5%; 7964/8804) completed the six-month duration. Participants enrolled into the HeartHealth program had a mean age of 58.6 years, 60.3% were male, and 62.4% were recruited from an outpatient clinic setting. A total of 851058 SMS messages were sent, with 99.41% delivered successfully. 3533 (44.4% of program completers) participants completed the post-intervention survey, and four HeartHealth staff members participated in a focus group discussion. Among the participants who completed the survey, 60.5% reported that the program improved the healthiness of their diet, 53.6% reported improved physical activity levels, and 56.1% reported that it helped remind them to take their medications. Content analysis of participant feedback identified that the program was effective in prompting participants to change their diet, providing emotional support, reminding them of the importance of behaviour change, improving their confidence in managing their health, and keeping participants focused. Key barriers included limited personalisation, language options, and SMS scheduling flexibility. Recommended adaptations focused on enhancing personalisation, greater engagement by local clinical teams and expanding program dissemination.</p><p><strong>Conclusions: </strong>The program had a broad reach, translated to improved patient-reported health behaviours, provided participants with needed support at low cost and low resource requirements. This analysis highlights the successful implementation and scalability of the HeartHealth program and provides key learnings for health systems who are looking to implement similar programs in the future.</p><p><strong>Clinicaltrial: </strong></p>","PeriodicalId":14706,"journal":{"name":"JMIR Cardio","volume":" ","pages":""},"PeriodicalIF":2.2,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12978537/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146029530","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Technologies, Clinical Applications, and Implementation Barriers of Digital Twins in Precision Cardiology: Systematic Review.","authors":"Fatemeh Sarani Rad, Ehsan Bitaraf, Maryam Jafarpour, Juan Li","doi":"10.2196/78499","DOIUrl":"10.2196/78499","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Digital twin systems are emerging as promising tools in precision cardiology, enabling dynamic, patient-specific simulations to support diagnosis, risk assessment, and treatment planning. However, the current landscape of cardiovascular digital twin development, validation, and implementation remains fragmented, with substantial variability in modeling approaches, data use, and reporting practices.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Objective: &lt;/strong&gt;This systematic review aims to synthesize the current state of cardiovascular digital twin research by addressing 11 research questions spanning modeling technologies, data infrastructure, clinical applications, clinical impact, implementation barriers, and ethical considerations.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;We systematically searched 5 databases (PubMed, Scopus, Web of Science, IEEE Xplore, and Google Scholar) and screened 330 records. Forty-two original studies met the predefined eligibility criteria and were included. Data extraction was guided by 11 thematic research questions. Mechanistic and artificial intelligence (AI) or machine learning (ML) modeling strategies, data modalities, visualization formats, clinical use cases, reported impacts, limitations, and ethical or legal issues were coded and summarized. Risk of bias was evaluated using a custom checklist for modeling studies, the Prediction Model Risk of Bias Assessment Tool (PROBAST) for prediction models, and the Risk of Bias in Non-Randomized Studies - of Interventions (ROBINS-I) for observational studies.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;Most digital twins (29/42, 69%) relied on mechanistic models, while hybrid mechanistic-data-driven approaches and purely data-driven designs were less frequent (13/42, 31%). Only 18 studies explicitly described ML or AI, most often deep learning, Bayesian methods, or optimization algorithms. Personalization depended primarily on imaging (32/42, 76%) and electrocardiography or other electrical signals (18/42, 43%). Visualization was dominated (41/42, 98%) by static figures and anatomical snapshots. Clinically, digital twins were most commonly applied to therapy planning, risk prediction, and monitoring. Reported benefits focused on improved decision-making and therapy-related impacts, with occasional (8/42, 19%) reports of increased accuracy or faster diagnosis, but there was limited evidence for downstream improvements in patient outcomes. Key barriers included strong model assumptions and simplifications; high computational cost; data quality and availability constraints; limited external validation; and challenges in real-time performance, workflow integration, and usability. Explicit discussion of ethical, legal, or governance issues was rare (7/42, 17%).&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Conclusions: &lt;/strong&gt;Cardiovascular digital twins show substantial potential to advance precision cardiology by linking personalized physiological models with clinical decision support, particularly for therapy planni","PeriodicalId":14706,"journal":{"name":"JMIR Cardio","volume":"10 ","pages":"e78499"},"PeriodicalIF":2.2,"publicationDate":"2026-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12782626/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145933275","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predicting Atrial Fibrillation Ablation Outcomes: Machine Learning Model Development and Validation Using a Large Administrative Claims Database.","authors":"Yijun Liu, Mustapha Oloko-Oba, Kathryn A Wood, Michael S Lloyd, Joyce C Ho, Vicki Stover Hertzberg","doi":"10.2196/77380","DOIUrl":"10.2196/77380","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Atrial fibrillation (AF) ablation is an effective treatment for reducing episodes and improving quality of life in patients with AF. However, long-term AF-free rates after AF ablation are inconsistent across the population, ranging from 50% to 75%. Patient selection relies on individual clinical assessment, highlighting a critical gap in population-level predictive analytics. While existing risk scores (eg, CHADS₂ [congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, and stroke], CHA₂DS₂-VASc [congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, stroke, vascular disease, age, and sex category], CAAP-AF [coronary artery disease, left atrial diameter, age, AF, antiarrhythmic drugs, and female sex category]) have been applied to predict AF ablation outcomes, their performance in administrative claims data remains unclear. Leveraging large administrative claims databases represents an opportunity to develop standardized, scalable prediction models that could inform population health management and resource allocation at a national level.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Objective: &lt;/strong&gt;This study utilizes machine learning (ML) models on claims data to explore if integrating International Classification of Diseases (ICD) billing codes outperforms traditional stroke and AF risk scores in predicting 1-year AF ablation outcomes.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;We analyzed claims data from the Merative MarketScan Research Medicare database (2013-2020) to identify 14,521 patients who underwent AF ablation. To predict 1-year AF-free outcomes, we developed logistic regression and extreme gradient boosting (XGBoost) models using demographic characteristics, comorbidity indices, and ICD diagnostic codes from the 2 years preceding ablation. Model predictions were compared with claims-based implementations of established risk scores-CHADS2, CHA2DS2-VASc, and a modified CAAP-AF (without left atrial diameter and the number of failed antiarrhythmic drugs). The ML models were also assessed on subgroups of patients with paroxysmal AF, persistent AF, and both AF and atrial flutter from October 2015 onward.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;Among 14,521 patients (mean age 71.5, SD 5.31 y; n=5800, 39.94% female), AF ablation success occurred in 54.01% (n=7843). XGBoost achieved areas under the receiver operating characteristic curve (AUCs) of 0.528, 0.521, and 0.529 for the whole, female, and male AF ablation groups, respectively, and better discrimination than CHADS2, CHA2DS2-VASc, and the modified CAAP-AF in all AF ablation groups (whole population, female, and male). While CHA2DS2-VASc and the modified CAAP-AF showed higher recall (&gt;0.798), their precision (&lt;0.540) was lower than XGBoost (0.552-0.556). In subgroup analyses of International Classification of Disease, Tenth Revision (ICD-10) patients (n=7646), the models incorporating ICD codes demonstrated better performance than those using only demographic and","PeriodicalId":14706,"journal":{"name":"JMIR Cardio","volume":"9 ","pages":"e77380"},"PeriodicalIF":2.2,"publicationDate":"2025-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12755845/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145878268","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Self-Reported Acceptance of a Wearable Activity Monitor in Persons With Stroke: Usability Study.","authors":"Jamie Nam, Grace C Bellinger, Junyao Li, Margaret A French, Ryan T Roemmich","doi":"10.2196/70007","DOIUrl":"10.2196/70007","url":null,"abstract":"<p><strong>Background: </strong>Wearable activity monitors offer clinicians and researchers accessible, scalable, and cost-effective tools for continuous remote monitoring of functional status. These technologies can complement traditional clinical outcome measures by providing detailed, minute-by-minute, remotely collected data on a wide array of biometrics, including physical activity and heart rate. There is significant potential for the use of these devices in rehabilitation after stroke if individuals will wear and use the devices; however, the acceptance of these devices by persons with stroke is not well understood.</p><p><strong>Objective: </strong>This study investigated the self-reported acceptance of a commercially available, wrist-worn wearable activity monitor (the Fitbit Inspire 2; Fitbit Inc) for remote monitoring of physical activity and heart rate in persons with stroke. We also assessed relationships between reported acceptance and adherence to wearing the device.</p><p><strong>Methods: </strong>Sixty-five participants with stroke wore a Fitbit Inspire 2 for 3 months, at which point we assessed acceptance using the Technology Acceptance Questionnaire (TAQ), which includes 7 dimensions: perceived usefulness, perceived ease of use, equipment characteristics, privacy concerns, perceived risks, facilitating conditions, and subjective norm. We then performed Spearman correlations to assess relationships between acceptance and adherence to device wear, calculated as both the percentage of daily wear time and the percentage of valid days the device was worn during the 3 weeks preceding TAQ administration.</p><p><strong>Results: </strong>Most participants reported generally agreeable responses, with high overall total TAQ scores across all 7 dimensions, indicating strong acceptance of the device; \"Agree\" was the median response to 29 of the 31 TAQ statements. Participants generally found the device beneficial for their health, efficient for monitoring, easy to use and to don and doff, and unintrusive to daily life. However, participant responses on the TAQ did not show significant positive correlations with measures of actual device wear time (all P>.05).</p><p><strong>Conclusions: </strong>This study demonstrates generally high self-reported acceptance of the Fitbit Inspire 2 among persons with stroke. Participants reported general agreement across all 7 TAQ dimensions, with minimal concerns interpreted as being directly relatable to poststroke motor impairment (eg, donning and doffing the device, using it independently). However, the high self-reported acceptance scores did not correlate positively with measures of real-world device wear. Accordingly, it should not be assumed that persons with stroke will adhere to wearing these devices simply because they report high acceptability.</p>","PeriodicalId":14706,"journal":{"name":"JMIR Cardio","volume":"9 ","pages":"e70007"},"PeriodicalIF":2.2,"publicationDate":"2025-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12726820/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145819295","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Explainable Logistic Regression for Heart Disease Risk Prediction in Community and Clinical Populations: Development and External Validation Study.","authors":"Peihua Tong, Hui Hu, Ling Tong","doi":"10.2196/82462","DOIUrl":"https://doi.org/10.2196/82462","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Heart disease is a leading cause of morbidity and mortality worldwide. Although machine learning models can achieve strong predictive performance, their limited interpretability hampers clinical adoption. Logistic regression is transparent but is often perceived as less accurate than complex ensemble models.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Objective: &lt;/strong&gt;To develop an explainable logistic regression model (SHAP-LR) for heart disease risk prediction using routinely available clinical variables and to evaluate its performance across community survey data, public clinical datasets, and a hospital cohort, in comparison with machine learning models and the Framingham Risk Score (FRS).&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;We used the 2015 Behavioral Risk Factor Surveillance System (BRFSS; 253,680 adults, 9.4% with self-reported heart disease) for model development. To benchmark machine learning methods, we trained baseline models on the full UCI Heart Disease dataset (n=920) and the Statlog Heart Disease dataset (n=270). The final SHAP-LR model itself was developed exclusively on BRFSS data. External validation of SHAP-LR was performed on the Cleveland subset of the UCI Heart Disease database (n=303), where SHAP-LR was benchmarked against FRS for discrimination and calibration.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;In BRFSS, older age and cardiometabolic risk factors were strongly associated with heart disease. Across the UCI, Statlog, and BRFSS datasets, SHAP-LR achieved AUROCs of approximately 0.73, 0.64, and 0.80, with performance comparable to or slightly better than more complex tree-based models. In the external cohort, SHAP-LR showed overall similar discrimination to FRS. Apparent calibration, as judged by Brier scores and calibration plots, was more favorable for SHAP-LR in this high-prevalence hospital sample, but this likely reflects the use of class-weighted training in BRFSS and the mismatch between a prevalence model and a 10-year incidence risk score; these calibration differences should therefore be interpreted with caution. Subgroup analyses indicated that FRS achieved higher AUROC than SHAP-LR in some high-risk groups, including patients with diabetes or hypertension. In the BRFSS test set, the corrected SHAP-LR integer score defined three strata with observed event rates of approximately 1.1%, 4.1%, and 17.1%; mean predicted probabilities were approximately 9.3%, 26.2%, and 60.7%, indicating effective risk ranking but substantial overestimation of absolute risk in the low-risk group.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Conclusions: &lt;/strong&gt;We developed and validated an explainable logistic regression model for heart disease risk prediction that balances predictive performance and transparency. By modeling age as a continuous predictor, comparing against multiple machine learning models, and using FRS as an external benchmark in a hospital cohort, SHAP-LR demonstrates a simple, interpretable framework for prevalent heart disease risk prediction in ","PeriodicalId":14706,"journal":{"name":"JMIR Cardio","volume":" ","pages":""},"PeriodicalIF":2.2,"publicationDate":"2025-12-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145804434","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}