Frontiers in digital health最新文献

China is one of the countries with a heavy burden of chronic diseases in the world. The number of people with chronic diseases and the number of deaths show a continuous, rapid growth trend. In recent years, China has actively promoted the innovation of digital management models for chronic diseases in grassroots communities, improved the allocation of community resources, clarified service specifications, strengthened information construction, and established a supervision and evaluation system. Digital management measures for chronic diseases in grassroots social areas include building a digital health management platform, promoting the popularization and application of intelligent detection equipment, and carrying out mobile medical and telemedicine services. However, it also faces some challenges, such as data security, privacy protection, uneven technology application, and patient acceptability. China should propose targeted measures to address these challenges, further promote the digital management of chronic diseases in grassroots communities, and protect the lives and health of people.

Background: Sierra Leone has advanced its digital health agenda. However, digital literacy among frontline health workers remains low, with over 82% reporting limited confidence in using digital tools. The health workforce also recorded the lowest digital health maturity score among all enablers in the WHO Global Digital Health Monitor, underscoring the need for workforce upskilling as a foundation for digital transformation.

Objective: This paper describes the design, implementation, and outcomes of a foundational digital literacy training program for frontline health workers under the Digital Innovation in Pandemic Control (DIPC) project, aimed at improving readiness for digital tool adoption.

Methods: A training needs analysis (TNA) aligned skill gaps with the competencies required for using the electronic Stock Management Tool (eSMT). Training modules were adapted from the European Commission's DigComp framework, contextualized for Sierra Leone, and delivered through a blended learning model. Post-training competency gains were assessed to determine effectiveness.

Results implementation: Among 150 trained health workers, "high understanding" in basic computer literacy increased from 7.1% to 72.2%, while "low understanding" dropped from 65.9% to 9.2%. For computer troubleshooting skills, "high understanding" rose from 4.4% to 73.8%. Both courses showed large effect sizes (Cohen's d = 1.3-2.1), indicating substantial learning gains.

Conclusions: Systematic digital literacy training, grounded in competency frameworks and contextual design, can substantially improve digital readiness among frontline health workers. Such interventions are essential foundations for sustainable digital transformation in health systems.

Background: Chronic obstructive pulmonary disease (COPD) exacerbations and hospital readmissions are major drivers of morbidity, mortality, and healthcare costs. Artificial intelligence and machine learning (AI/ML) approaches have been applied to predict these events, but their pooled performance and methodological rigor remain unclear.

Methods: Following PRISMA 2020 guidelines, we conducted a systematic review and meta-analysis of peer-reviewed studies developing or validating AI/ML models for predicting acute exacerbations of COPD (AECOPD) or hospital readmissions. Databases (PubMed, IEEE Xplore, Cochrane Library, Semantic Scholar) were searched to 2025. Eligible designs included retrospective and prospective cohorts, randomized trials with embedded prediction, and case-control studies. Study quality was assessed using PROBAST, and evidence certainty with GRADE. Random-effects models pooled area under the ROC curve (AUC); subgroup analyses compared AECOPD vs. readmission outcomes and internal vs. external validation.

Results: Thirteen studies were included, with sample sizes ranging from 110 to 113,786 patients. Most were retrospective cohorts using EHRs or claims data, while two used prospective or trial-based data. Models applied diverse algorithms, including random forests, gradient boosting, neural networks, and ensemble pipelines. The pooled AUC across all studies was 0.77 (95% CI: 0.74-0.80), with very high heterogeneity (I² = 99.5%). Subgroup analyses showed similar performance for AECOPD prediction (AUC = 0.77; I² = 98.9%) and readmission prediction (AUC = 0.73; I² = 19.8%). Externally validated models (n = 4) achieved higher accuracy (AUC = 0.82) than internally validated models (AUC = 0.76), although differences were not statistically significant. Risk of bias was moderate to serious in 69% of studies, mainly due to incomplete reporting and overfitting.

Conclusion: AI/ML models demonstrate moderate-to-high discriminatory accuracy in predicting COPD exacerbations and readmissions, with pooled AUCs of 0.73-0.77. However, high heterogeneity, limited external validation, and frequent methodological concerns restrict generalizability. Standardized reporting frameworks (TRIPOD-AI, PROBAST-AI), rigorous external validations, and prospective implementation studies are needed to translate these promising tools into clinical practice.

Background: The rapid aging of populations poses major challenges to health and social care systems. Supporting older adults in managing chronic conditions while promoting independence and quality of life requires innovative approaches that extend beyond senior institutional care. Telehealth has emerged as a promising approach to enhance access, continuity, and patient engagement. However, evidence regarding its effectiveness and best practices remains fragmented.

Objectives: This systematic review aimed to synthesize current evidence on telehealth interventions for adults aged 65 years and older, focusing on their effects on health outcomes, quality of life, and well-being.

Methods: A search across three databases in the last five years identified 37 eligible studies, and data analysis was guided by a comprehensive taxonomy. Interventions were diverse, spanning disease management, rehabilitation, health promotion, clinical decision support, and psychological support.

Results: Reported benefits included improved physical function, better chronic disease control, greater health knowledge, and reductions in avoidable hospitalizations. Video-based programs showed greater effectiveness, while telephone-only interventions were most useful when combined with remote monitoring. Adherence was strengthened by professional guidance, caregiver support, and real-time feedback.

Discussion: Despite encouraging findings, evidence remains inconsistent regarding quality-of-life outcomes, cost-effectiveness, and scalability across populations, with many studies limited by small samples, short duration, and methodological heterogeneity. Telehealth holds the potential to complement traditional care for older adults across multiple clinical domains, and future research must adopt consistent and comprehensive reporting practices to strengthen decision-making and ensure that this pathway evolves with patients' needs.

Systematic review registration: PROSPERO CRD420251072656.

Adolescent idiopathic scoliosis (AIS) is traditionally assessed and classified using radiographic methods that rely on Cobb angle measurements and qualitative curve modifiers, exposing patients to repeated radiation and offering limited sensitivity to subtle three-dimensional (3D) deformities. We developed ScolioClass, a non-invasive, data-driven classification tool that harnesses 3D optical surface scanning and continuous indices, capturing curvature severity, directionality, and sagittal balance, to evaluate spinal deformities in 94 patients with AIS. By comparing ScolioClass descriptions with the established Lenke classification, we observed a statistically significant association (χ ² ≈ 29.0, df = 6, p < 0.001) with 72.3% overall agreement. A significant association was also found between sagittal modifiers and ScolioClass kyphosis-lordosis categories (χ ² ≈ 48.4, df = 3, p < 0.0001) with 68.1% agreement. Notably, ScolioClass detected mild curves and lordotic patterns that were often overlooked by Lenke criteria. These findings demonstrate that ScolioClass provides radiation-free, quantitative 3D assessment of AIS with potential for automated analysis and individualized treatment planning. Further validation is warranted for clinical integration.

Wearable technologies mark a transition in healthcare evolution, from paternalistic (Healthcare 1.0) to reactive (Healthcare 2.0), proactive (Healthcare 3.0), and data-integrated care (Healthcare 4.0). The next stage, Healthcare 5.0, envisions the seamless integration of human and pet health data, fostering a more holistic approach to disease prevention and management. In this viewpoint, we explore the disruptive potential of integrating health monitoring between humans and pets through wearable technology, highlighting the interconnected nature of human-pet health. We examine the parallel evolution of human and pet health monitoring, assessing current technologies and their potential to enhance both fields. We discuss that wearable technologies not only improve chronic disease management but also enable early detection of zoonotic and emerging diseases. Additionally, we emphasize the potential re-usability of human wearable devices for pets, outlining the associated technical challenges. This can lower costs and accelerate adoption, offering mutual benefits for both domains. We address the need for an integrated, linked platform that enables real-time data analysis. Data integration ultimately results in better diagnostic accuracy, optimized treatment plans, and enhanced quality of life for humans and pets. Re-purposing wearables for human-pet health monitoring enables real-time data collection, predictive analytics, and prevention to accelerate the implementation of Healthcare 5.0.

Introduction: Use of digital health technology can improve patient health outcomes; however, not all patients have the knowledge and skills to download a health app and access a patient portal. Providing digital onboarding support to hospitalized patients has potential to overcome some barriers to accessing needed education in the community, including both having the time and a location to receive education. To address this, our team developed the Jefferson Digital Onboarding Taskforce (JeffDOT), a group of staff and students who approach hospitalized patients and provide one-on-one teaching on how to sign up for and use a patient portal.

Methods and materials: This descriptive study examined the implementation and preliminary outcomes of JeffDOT. We collected patient demographics and assessed health literacy, digital health readiness, and empowerment using the patient portal after patients received individualized support with portal enrollment.

Results: We enrolled 343 hospitalized patients from a large academic medical center in the U.S. in their patient portal. Almost half of the sample (49%) was older than 55 years, 56% were male, 34% were Black, and 19% spoke Spanish at home. After receiving individualized support from the JeffDOT team, the majority of patients (84%) reported that they felt empowered to use the portal and almost half reported that they would be very interested in additional basic computer skills training if offered by the hospital.

Discussion: Our findings suggest that supporting hospitalized patients with enrollment into a health portal using a primarily student, volunteer-staffed model is feasible and acceptable to patients. Future research should focus on the impact of JeffDOT on patient outcomes and health behaviors.

Background: Poor access to antenatal care (ANC), skilled delivery, and postnatal checks within 48 h of delivery are linked to adverse pregnancy outcomes. In Kenya, unequal use of these services has caused significant regional disparities, with 15 out of 47 counties being high priority.

Objectives: To evaluate the effectiveness of a digital health solution to improve maternal and newborn health (MNH) uptake.

Methods: From July 2017 to March 2019, we implemented an integrated community and hospital digital health information system, in ten health facilities and four community health units in Turkana County, Northern Kenya. We used a non-randomized before-after household survey. We assessed the proportion of mothers attending at least four antenatal visits, receiving skilled delivery, and receiving postnatal checks within 48 h at baseline and 12 months post-intervention. Statistical comparisons included p-values and 95% confidence intervals, accounting for clustering at the CHU level. These findings were compared with data from the Kenya Health Information System for the study subcounty and Turkana County.

Results: Among a catchment population of 4,300 women of reproductive age (WRA), 692 and 608 women were interviewed at baseline and endline, respectively. STONE-HMIS® led to 5%, 23%, and 16% improvements in 4th antenatal care visits, skilled delivery, and postnatal checks within 48 h, respectively. For the same period, subcounty and county data showed that 57.7% and 65.8% of WRA attended at least 4 ANC visits, 39.5% and 67.8% delivered with skilled assistance, and 23.5% and 24% had postnatal checks.

Conclusions: Integrating digital health systems at provider and community levels, aligned with health system priorities, showed marked improvements MNH indicators in an underserved, remote area.

The increasing digitalization of health systems is accelerating the transition towards a new era of data-driven, evidence-based care. This has profound implications for clinical practice, performance evaluation, policy making and biomedical research. At the heart of this transformation lies a healthcare data warehouse (DW), which functions as a critical infrastructure for aggregating, standardizing, and analyzing diverse clinical and administrative data. When well-designed and implemented, DWs provide clinicians with timely access to comprehensive, longitudinal patient data, enabling more informed decision-making, enhancing care quality, and improving outcomes. For researchers, these repositories offer opportunities for population-level analytics, predictive modeling, and large-scale health service research, enabling insights into disease patterns, healthcare utilization, and system inefficiencies. Centralizing clinical and administrative data in a DW allows for more frequent, nuanced analyses, increasing the precision and responsiveness of care. However, developing an effective DW requires careful consideration of system architecture, data governance, and interoperability. These foundational components support the robust ETL/ELT frameworks that ensure data quality, consistency, and readiness for analysis across diverse and evolving data streams. Beyond supporting individual patient care, DWs act as essential drivers of scalable research, operational efficiency, and evidence-based health policy. Their successful implementation marks a pivotal step toward achieving personalized, high-quality, and cost-effective healthcare in the digital transformation age. This paper reviews the existing literature to outline the process of building and implementing a data warehouse, introducing real-world disease-specific applications. BiotherDW connects theoretical frameworks with practical healthcare applications by demonstrating how traditional data warehouse design can be adapted for national-scale digital health infrastructures.