BackgroundPain is a significant risk factor for depression among older adults. While prior studies suggest that internet media may improve mental well-being, it remains unclear whether such media can reduce pain-related depression.ObjectivesThis five-year longitudinal study explores the potential moderating effect of internet media on the relationship between pain and depression among older adults.MethodsParticipants were sourced from 2018, 2020, and 2023 waves of the China Longitudinal Aging Social Survey, and this study utilized 3240 “person-year” observations from 1080 respondents. An individual fixed effects model was employed. The presence of pain, depression (measured by the CES-D scale), and media preference (measured by comparing internet and traditional media use frequency) were assessed. Subgroup heterogeneity was also explored.ResultsThe findings revealed that media preference significantly moderated the relationship between pain and depression among older adults (β = −0.725, p < .01). Compared with traditional media, internet media was more effective in alleviating depression in individuals experiencing pain. The engagement breadth of internet media also exhibited a buffering effect. Heterogeneity analysis further illustrated that the beneficial effects of internet media were more pronounced among older adults who were less educated (β = −0.865, p < .01) and retired (β = −0.887, p < .01).ConclusionThis study enhances the understanding of the theoretical and practical aspects of internet media's moderating role in depression among older adults. It also highlights heterogeneous effects in vulnerable subpopulations. The findings offer insights for developing non-pharmacological interventions to address depression associated with pain, contributing to promoting mental health in the aging population.

Uterine contraction is a meaningful indicator for labor onset and appropriate hospital admission. Inaccurate self-assessment may lead to premature admission, unnecessary interventions, and higher healthcare resource use. Traditional monitoring devices have limited portability and comfort, restricting home-based use. This study developed and validated a wearable system integrating flexible sensors, a data acquisition platform, and machine learning models to monitor uterine contractions and identify labor onset, focusing on late pregnancy and the pre-labor period. A flexible sensor-based device was developed and validated against hospital toco. Contraction data from 82 participants (104 recordings) were preprocessed and segmented, and features were extracted for model training. Hospital admission was classified into recommended admission (RA), deferred admission (DA), and selective admission (SA). Several ML models were trained and evaluated via 10-fold stratified cross-validation using accuracy, precision, recall, F 1-score, and area under the curve. Shapley Additive Explanations (SHAP) analysis interpreted feature contributions. A total of 82 participants were enrolled, and 104 uterine contraction recordings were collected, ranging from 10 to 70 min (mean 20.3). Two hundred and seventy-seven processed segments were obtained for analysis. Contraction signals were generally consistent with toco measurements (r = 0.85-0.95). XGBoost achieved accuracy of 0.87 for RA classification, and SHAP identified kurtosis, signal energy area, and standard deviation as key features. The system enabled accurate monitoring of uterine contractions, improved estimation of hospital admission timing, reduced premature admission risk, and demonstrated high wearability, offering a feasible solution for home obstetric monitoring.

In blended care therapy, clients and therapists work together in synchronous psychotherapy sessions and asynchronous guided practice sessions, in which therapists assign digital tools (e.g., video lessons, written exercises) and provide guidance. This model of care is effective at scale, but the therapist experience of delivering this care has not been thoroughly examined. This study explores the therapist experience of using digital tools in blended care therapy delivered via telehealth and the perceived benefits and challenges of working within this model. Therapists (n 276, including marriage and family therapists, professional counselors, social workers, and psychologists) were invited to complete a survey on how the digital components of the therapy platform impact care delivery. Twelve participants completed a semi-structured interview about their experience using digital tools while delivering care. Transcripts were analyzed using thematic analytic methods. The majority of survey participants reported that digital tools had positive effects on their mental health care delivery. Nearly all respondents (93.68%) reported learning new clinical skills from the tools. Three primary themes were generated through thematic analysis. First, we summarize ways in which blended care impacts therapeutic relationships. Then, we describe how the structure of digital tools supports clients and therapists to develop and practice skills. Finally, we discuss how delivering blended care impacts therapist time and energy. Results provide insights into the experience of delivering blended care therapy. Well-integrated digital tools appear to bring unique value to therapists that goes beyond their past experiences delivering care in traditional systems.

The objective of this retrospective study is to compare three learning approaches for blood transfusion (BT) prediction after intensive care unit (ICU) admission: local learning (LL), federated learning (FL), and centralized learning (CL) across five ICU databases (eICU Collaborative Research Database, Medical Information Mart for Intensive Care IV, High-Resolution Intensive Care Unit Dataset, Amsterdam University Medical Center Database, University Hospital of Augsburg). As machine learning model we used XGBoost and included 15 clinical variables. The prediction task consists of a 3-h observation window, followed by a 2-h prediction window. We evaluated the models using internal and external validation with area under the receiver-operator curve, area under the precision-recall curve, PPV, Brier score and F1 score. CL consistently outperformed FL and LL in both internal validation (AUPRC range: 0.73-0.95 (CL) vs 0.63-0.96 (FL) and 0.69-0.96 (LL)) and external validation (AUPRC range: 0.61-0.89 (CL) vs 0.45-0.91 (FL) and 0.37-0.90 (LL)). FL showed variable performance across datasets. The complexity of the multivariable clinical prediction of BTs may create substantial challenges for FL effectiveness, particularly under high data heterogeneity conditions that are common in healthcare.

Large language models have a huge positive impact on various disciplines, including healthcare. As family caregivers are an essential part of the healthcare system, they need support and can benefit from the technology. However, there is no consensus on reliable and valid measures to evaluate large language models. This study aims to review the literature on the evaluation measures of large language models for caregivers. We conducted a scoping review guided by Arksey and O'Malley methodology and the PRISMA-ScR checklist. A literature search on PubMed, EMBASE, CINAHL, and PsycINFO, from 2018 through July 2024, was carried out. An additional rapid review was conducted for the recent literature update from July 2024 through November 2025. All 10 final publications that met the inclusion criteria out of 1812 focused on ChatGPT, whereas three of them also addressed other large language models, such as Google Bard and Bing AI. The most commonly assessed core conceptual components of evaluation measures were accuracy, reliability, readability, and comprehensiveness. Overall, the included studies reported that large language models' responses were somewhat accurate and reliable and mixed results in readability and comprehensiveness. The final 14 publications from a rapid review offered additional evidence on ChatGPT-centrism. This review provides a comprehensive overview of the measures for evaluating large language models and highlights the need for their improvement using reliable and valid measures. The findings guide the direction of future research and practice to maximize the benefits through continuous quality improvement.

This study aimed to evaluate the current research progress and future research directions of artificial intelligence in spinal diseases through a bibliometric analysis. Publications regarding spinal diseases and artificial intelligence published from 2006 to 2025 were extracted from the Web of Science Core Collection (WOSCC). Subsequently, a bibliometric analysis of these publications was conducted using CiteSpace, VOSviewer, and Bibliometrix Online Analysis Platform. A total of 734 papers were included in the study. The annual publication numbers are on the rise. The USA (267 papers) and Beijing Jishuitan Hospital (24 papers) were identified as the most productive country and institution, respectively. Tian, Wei (25 papers) is the most productive author. "Spine" (49 publications) is the most productive journal. "Machine Learning" was the most cited keyword, with high-frequency keywords such as "Artificial Intelligence," "Robotic Surgery," "Deep learning," "Virtual Reality," and "predictive modeling" signaling hot topics in this field. There are increasingly many papers on artificial intelligence in spinal diseases. However, cooperation between institutions in various countries needs to be strengthened. In addition, this study summarized the research focus of artificial intelligence in spinal disorders as accurate diagnosis, robot-assisted surgery, and prognosis prediction, providing researchers with future research directions.

The rapid growth in eHealth has opened new possibilities for people with Parkinson's Disease (PD). However, given the complexity of PD symptoms, the successful adoption of eHealth likely depends on support from family caregivers. While previous research has explored the perspectives of people with PD, insights into family caregivers' perceptions of eHealth remain limited. To explore family caregivers' perceptions of eHealth in formal care and in supporting physical activity and exercise. A qualitative design was used. Interviews were conducted with 15 family caregivers (mean age 72 years, 12 women) from rural and urban Sweden. Data were analyzed using inductive qualitative content analysis. Three themes were formed. Being the driving force for home, health and technology described family caregivers' responsibility for managing technology, balancing confidence with concerns about added burden and emotional strain, and supporting their partner's engagement. A threat to or enabler of quality time reflected hopes that eHealth could save time and offer greater flexibility to schedule exercise and activities around daily routines, while also raising concerns about isolation and reduced social interaction if digital solutions replaced in-person contact. This theme also emphasized the need for accessible exercise solutions tailored to PD symptoms. Bridging the gaps: blending digital and in-person care highlighted reliance on online information due to limited professional guidance, appreciation for timely access to PD nurses, and a preference for integrated care models combining in-person care and digital follow-up. In conclusion, family caregivers highlighted both opportunities and challenges with eHealth in PD care, valuing its potential to ease daily responsibilities and enhance flexibility, while expressing concerns about added burden, reduced social interaction, and limited professional guidance.

This study aimed to systematically review randomized controlled trials evaluating the effectiveness of web- and app-based health management interventions for patients at high risk of stroke, including those with cardiovascular disease, type 2 diabetes, or obesity. Clinical outcomes (blood pressure and glycated hemoglobin), self-management knowledge, health behaviors, and quality of life were examined. Following PRISMA guidelines, we searched PubMed, Web of Science, CINAHL, and the Cochrane Library for randomized controlled trials published between January 2010 and December 2024. Search terms included combinations of "Hypertension," "Diabetes Mellitus," "Myocardial Infarction," "Obesity," "Mobile Applications," and "Web-Based Intervention." Studies that provided sufficient data for quantitative synthesis were included in the meta-analysis. Pooled effect sizes were calculated using RevMan 5.4, and subgroup analyses were conducted based on intervention duration. Thirty-five studies met the inclusion criteria. Meta-analysis revealed that web- and app-based interventions significantly improved blood pressure, glycated hemoglobin level, and medication adherence among individuals at high risk of stroke. These interventions also enhanced quality of life and increased self-management knowledge and behaviors. Subgroup analyses of web- or app-based educational programs indicated that clinical indicators improved with interventions lasting 3 months, whereas self-management knowledge, behaviors, and depression symptoms showed significant improvements with 6-month interventions. This study demonstrates that web- and app-based health management interventions yield improvements in clinical, psychological, and behavioral outcomes for patients at high risk of stroke. These findings provide evidence for the development and implementation of tailored digital health interventions for effective stroke prevention and chronic disease management.

IntroductionPhotoplethysmography (PPG) based heart rate (HR) monitoring supports continuous assessment of physical activity intensity, but device generated HR values remain prone to motion artefacts. This study validated a novel machine learning procedure that detects artefacts directly in PPG derived HR data, rather than in raw waveforms that are unavailable in commercial devices.MethodsIn this prospective study, 149 participants (46 following cardiac rehabilitation, 103 healthy) wore a PPG-based wrist device and a reference chest strap for 12 weeks. Prior testing defined participants as “PPG-compatible” (i.e., HR error <10% during ≥70% of training data). Three quarters trained artefact and activity detection models, one quarter served as the test population. To balance artefact removal and activity preservation, multiple probability thresholds were combined to reject unreliable HR values, remove or interpolate using adjacent reliable values. The Antwerp Activity Index (AAI), an HR-based PA score, was calculated from the resulting PPG and reference data to evaluate threshold combinations.ResultsSeventy-eight participants (53.8%) were PPG-compatible, yielding over 5 million datapoints, with 75% (n = 58; 4,144,654 datapoints) used for training and 25% (n = 20; 992,180 datapoints) for testing. The artefact model detected artefacts with a sensitivity of 58.5% in daily life and 76.6% during exercise, while limiting incorrect removals (74.9–91.3% specificity). Rejecting data for AAI-calculation when artefact probability was >50% and activity probability <70% produced the highest agreement with reference AAI (Pearson’s r = .89–.96).ConclusionsThis new machine learning-based procedure, which operates on device-generated HR rather than PPG waveforms, effectively removes artefacts while preserving key activity signals, ensuring clinically relevant PA assessment in daily life.

The application of digital technology in jaw reconstruction is clinically significant as it enhances surgical precision and patient outcomes. This study aimed to conduct a comprehensive bibliometric analysis to explore global research trends, hotspots, and future directions in the application of digital technology for jaw reconstruction. This study employs a bibliometric analysis design, utilizing a retrospective approach. The data were sourced from the Web of Science Core Collection (WoSCC) database, focusing on publications related to digital technology in jaw reconstruction from 1995 to 2024. The final sample includes 1069 articles authored by 4781 authors from 3081 institutions. Data analyses were conducted using VOSviewer, CiteSpace, and the R package "Bibliometrix." The analysis revealed a total of 1069 articles, with publication trends showing a peak in 2022. China produced the most articles, while the United States led in citations. The most productive institutions were based in Italy, Germany, and China, and key contributing authors included Marchetti Claudio and Ciocca Leonardo. Journal of Cranio-Maxillofacial Surgery had the highest H-index among journals. Keyword co-occurrence revealed hotspots like "implants" and "computer-aided design," while burst terms such as "virtual surgical planning" and "3D printing" signaled emerging frontiers. This bibliometric analysis examined the global research landscape on digital technology in jaw reconstruction, identifying key developments such as rapid prototyping, virtual surgical planning, and 3D printing. Future research should focus on integrating these technologies into clinical workflows, reducing costs, improving training, and addressing regulatory challenges.