Objectives Given the persistent gap in routine care between functional assessment and individualised therapeutic exercise prescription among the approximately 1.3 billion people worldwide living with significant functional limitations, this systematic review aimed to identify and characterise named digital platforms and technology enabled systems that integrate functional assessment, therapeutic exercise prescription, and longitudinal monitoring, describing their integration models and the design features reported alongside adherence and functional outcomes. Clinical, measurement, and implementation outcomes are summarised as complementary evidence within this framework. Methods We conducted a systematic review following PRISMA guidelines. The protocol was retrospectively registered in the Open Science Framework (OSF) on 25 May 2026 ( https://osf.io/sb87d ). Four electronic databases (PubMed, Web of Science, Scopus, and CINAHL) were searched up to 30 May 2025, with eligibility restricted to studies published in English, Portuguese, or Spanish. Inclusion criteria were defined using the PICOS framework. Data extraction included integration models, assessment–prescription linkage, intervention components and design features, clinical and functional outcomes, and intervention duration and follow-up. Risk of bias and methodological quality were assessed using the NHLBI Quality Assessment Tools appropriate to each study design. Results A total of 12 studies were included, covering 748 participants across diverse clinical and rehabilitation contexts plus 5 rehabilitation professionals in a usability pilot. The included platforms and systems ranged from directly linked assessment–prescription workflows combining assessment, prescription and monitoring to tools focused primarily on remote functional assessment or on digital support. Interventions most frequently involved mobile health applications, web-based platforms, and tele-rehabilitation programmes. Design features such as individualised exercise prescription, remote clinician monitoring or supervision, structured user feedback, and self-monitoring recurred across several platforms and systems, while goal setting, prompts and reminders, gamification, and chat or video communication with clinicians were reported less consistently. Selected intervention studies reported preliminary signals of possible functional or quality-of-life benefit, but findings for adherence, feasibility, and usability were mixed and context-dependent. Owing to heterogeneity of interventions and outcomes, results were synthesised narratively; methodological quality ranged from poor to good, with recurrent issues in sample-size justification, blinding of outcome assessors, and incomplete reporting of intervention exposure. Conclusions Named digital platforms and technology enabled systems that integrate functional assessment, therapeutic exercise prescription, and longitudinal monitoring are promising but still preliminary. Evidence to date provides preliminary signals of feasibility, acceptability, and possible short-to medium-term functional benefit in selected intervention studies, but remains heterogeneous and methodologically fragile. Large, multicentre randomised controlled trials with standardised outcome measures and longer follow-up periods are required to establish sustained benefit. For now, these tools should complement, not replace, in-person care, helping to address gaps in access and continuity.
Objective This study conducted a bibliometric analysis to evaluate the global trends and thematic evolution of artificial intelligence (AI) in human infectious diseases from 2000 to 2025. Methods A dual-database analysis was performed by retrieving relevant publications from the Science Citation Index Expanded edition of the Web of Science Core Collection and Scopus. After cross-database de-duplication and manual screening for thematic relevance, bibliometric analyses and visualizations of the 19,861 included articles were conducted using the Bibliometrix R package. Results A total of 19,861 articles were included, with publication output increasing sharply during the COVID-19 pandemic and peaking at 3,684 publications in 2022. China and the USA were the most productive countries based on corresponding-author country, and the USA had the highest betweenness centrality in the international collaboration network. Co-citation analysis identified four clusters covering foundational AI methods, COVID-19 imaging applications, and clinical and diagnostic evidence. Thematic analysis identified two major keyword communities: machine learning for infectious diseases and public health, and deep learning-based diagnostic imaging. Recent topics included large language models, explainable AI, and antimicrobial resistance. Conclusion AI-related research in human infectious diseases has expanded from exploratory computational applications to a prominent and increasingly diversified area of public health research. Future research should prioritize interpretable, externally validated, and equitably deployable models that can be responsibly integrated into clinical and public health workflows.
Background Patient narratives provide rich insight into emotional, relational, and decisional dimensions of chronic illness that are often absent from structured clinical records. However, qualitative analysis of such narratives remains labor-intensive and difficult to scale. Emerging advances in natural language processing (NLP) and large language models (LLMs) offer opportunities to augment, rather than replace, interpretive qualitative inquiry. Objective This study introduces and demonstrates a hybrid qualitative–computational framework for mapping emotional tone and decision-making influences within patient narratives in chronic illness contexts. Methods We conducted semi-structured interviews with 70 adults diagnosed with cancer (n=33) or diabetes (n=37). A dual-path analytic workflow combined investigator-led thematic coding with LLM-assisted zero-shot analysis. Embedding-based relational modeling was applied to map references to key decision-making agents (e.g., clinicians, family members, spiritual figures) within thematic segments. Sentiment analysis and exploratory clustering were used to characterize emotional variability across narratives. Computational outputs were interpreted comparatively and integrated with qualitative findings. Results Substantial agreement was observed between human and LLM-assisted thematic coding (Cohen’s κ=0.71). Relational mapping identified distinct decisional influence patterns across illness contexts, with oncology narratives emphasizing physician, family, and spiritual agents, and diabetes narratives centering self-management and healthcare continuity. Emotional tone demonstrated greater variability among oncology participants. Clustering analyses supported descriptive differentiation of narrative patterns without identifying statistically distinct subgroups. Conclusions This study proposes a scalable, AI-augmented framework for relational mapping of patient narratives, illustrating how computational tools can extend qualitative analysis while preserving contextual depth. By systematically characterizing emotional and decisional dynamics, such approaches may inform digital health research, patient-centered communication strategies, and future integration of narrative-informed signals into health systems.
Objective Inadequate continuity of care after hospital discharge may hinder stroke rehabilitation, while eHealth technologies offer opportunities to extend nursing support beyond hospital settings. This systematic review aimed to synthesize evidence on nurse-delivered eHealth interventions for post-stroke continuity of care, with a focus on technology types, nursing roles, intervention components, and associated health outcomes. Methods Peer-reviewed studies were searched in PubMed, Embase, Web of Science, CINAHL, CNKI, WANFANG, and VIP from inception to April 30, 2025. Methodological quality was assessed using the Mixed Methods Appraisal Tool (MMAT), and findings were synthesized thematically. Results A total of 17 studies met the inclusion criteria, including 10 randomized controlled trials, 4 quantitative non-randomized studies, 2 qualitative studies, and 1 mixed-methods study. Mobile applications were the most frequently used type of eHealth technology. Nurse involvement varied across studies: 10 interventions were nurse-led, whereas seven involved nurses as members of multidisciplinary teams. Nurse-delivered eHealth interventions may support secondary prevention, telerehabilitation, and individualized care. The most consistent favorable findings were observed for medication adherence and motor function or functional recovery; improvements were also reported in self-efficacy, blood pressure control, depressive symptoms, emotional adjustment, and social participation or reintegration. However, findings were not entirely consistent across studies and should be interpreted cautiously given differences in study design, intervention characteristics, and outcome measures. Conclusion Nurse-delivered eHealth interventions show potential to support continuity of care after stroke. Future high-quality, standardized studies with longer follow-up periods are needed to establish their effects, applicability, and implementation potential.
Objective To map the characteristics, applications, and reported outcomes of clinical decision support systems (CDSS) in maintenance hemodialysis (MHD) management and identify priorities for future development and implementation. Methods A comprehensive literature search was conducted in PubMed, Web of Science, Embase, IEEE Xplore, CINAHL, Cochrane Library, ProQuest, CNKI, Wanfang Data, ChineseBiomedical Literature Database (CBM), VIP Database from inception to September 2025. The review followed the Joanna Briggs Institute methodology for scoping reviews and was reported in accordance with the PRISMA-ScR guidelines. Studies describing the development, implementation, or evaluation of CDSS applied to maintenance hemodialysis management were included. Results Twenty-three studies were included. Applications covered five non-mutually exclusive areas: treatment and medication adjustment, clinical assessment and risk stratification, intradialytic safety monitoring, education and information support, and patient information management and longitudinal monitoring. Systems used knowledge-based rules, predictive models, and monitoring technologies, with deployment through electronic health records, standalone software, web-based platforms, or dedicated devices. Intended users included physicians, nurses, and other healthcare professionals. Reported benefits included improvements in selected clinical outcomes, decision support performance, and care efficiency; however, findings were inconsistent across studies. Evaluation methods and outcome measures were heterogeneous, and reporting of stakeholder involvement and sustained implementation was limited. Conclusion CDSS may support individualized decision-making and more efficient MHD care, but their clinical benefits and generalizability require further evaluation. Future development should strengthen workflow integration, user-centered design, and multidisciplinary involvement.
