Objective To map provider payment models and reimbursement policies for remote patient monitoring (RPM) across settings, and describe barriers, facilitators, and economic or operational consequences. Introduction RPM use is growing, yet reimbursement remains unsettled. Fragmented policies and misaligned payment incentives may limit scale-up and sustainability. Inclusion criteria Peer-reviewed sources on providers, payers or RPM recipients reported an RPM payment model, reimbursement policy or gap plus economic, operational or implementation findings. Grey sources provided RPM-specific payment, policy or coverage mechanics. Methods A JBI-aligned scoping review used database and four-strand grey-literature searches, dual independent title and abstract screening, Excel charting, ATLAS.ti coding, and MMAT appraisal of peer-reviewed studies to support interpretive caution. Results Fifty-six sources were included (34 peer-reviewed; 22 grey). Among peer-reviewed studies, fee-for-service predominated, often using management windows aligned with review frequency. Episode payments and shared-savings were less common and programme-specific. Across both source types, common barriers included coverage uncertainty, administrative workload, and misaligned incentives, while facilitators included value-oriented elements and transition funding. Economic findings were mixed, and utilisation outcomes were often unreported. Sources emphasised clear eligibility, supervision and frequency rules, brief documentation, and responsibility for devices, platforms, and connectivity. Three cross-cutting mechanisms concerned incentives and money flow, rules and billing gates, and workflow and workload. Conclusions Based on recurring patterns in the evidence, we propose prospective evaluation of a substitutive fee-for-service management window aligned with review frequency. Bundles and shared-savings may suit longitudinal monitoring but require attribution and data infrastructure. Public Interest Summary Remote patient monitoring lets clinicians track health between visits using data sent from home. We reviewed 56 sources to see how these services are paid for and what makes them easier or harder to run. Among peer-reviewed studies, fee-for-service was the most frequently reported payment model, often through monthly payments covering monitoring and review. Fewer studies described fixed-period payments or shared-savings contracts. Common hurdles were unclear rules, heavy paperwork, and incentives that do not align, for example when clinics pay while savings appear elsewhere. Helpful features were clear eligibility, short checklists, and temporary funds to start services. Our main takeaway: match payment to how monitoring works over time. One option for future evaluation is a monthly “management window” that replaces some clinic visits and states who pays for equipment and platforms. Other models may also be considered as data sharing improves. This may support more sustainable remote care.
Background Artificial intelligence (AI) is increasingly used in ophthalmic screening, but evidence regarding its effects beyond diagnostic performance remains heterogeneous. This systematic review evaluated AI-assisted ophthalmic screening across diagnostic performance, healthcare service-delivery, patient-process, and direct clinical patient outcomes. Methods PubMed, Scopus, and Web of Science were searched for eligible studies published between January 2015 and December 2025. Twenty-one studies met the inclusion criteria. Data were synthesized without meta-analysis because of substantial heterogeneity in study designs, AI applications, and outcome measures. Risk of bias and certainty of evidence were assessed using design-appropriate appraisal tools and GRADE principles. Results AI systems demonstrated high diagnostic performance, particularly for diabetic retinopathy screening. Evidence also suggested improvements in selected service-delivery outcomes, including workflow efficiency and specialist productivity, and patient-process outcomes, including screening completion, referral uptake, and follow-up adherence. Three randomized controlled trials provided evidence for selected process and service outcomes. Evidence was less mature for glaucoma and OCT-based retinal disease applications. Direct clinical patient outcomes, including visual acuity, disease progression, and vision loss, were rarely evaluated. Certainty of evidence varied across outcome domains. Conclusions AI-assisted ophthalmic screening demonstrates strong diagnostic performance and may improve selected service-delivery and patient-process outcomes. However, evidence for direct long-term clinical benefits and large-scale implementation remains limited. Context-specific validation, governance, economic evaluation, and prospective implementation studies are needed before widespread adoption. Public Interest Summary Artificial intelligence (AI) is increasingly being used to help detect eye diseases and support screening services. This systematic review examined 21 studies evaluating AI in ophthalmic care. The strongest evidence concerned diabetic retinopathy screening, where AI demonstrated high diagnostic performance. AI may also help improve how eye-care services are delivered by supporting workflow efficiency and specialist productivity. Some studies reported improvements in patient-process measures, such as completing screening, attending referrals, and adhering to follow-up. However, these measures should not be interpreted as evidence that AI directly improves patients' long-term eye health. Evidence showing improvements in visual acuity, disease progression, or prevention of vision loss remains limited. Further studies are needed to determine whether AI can be safely, equitably, and cost-effectively integrated into routine ophthalmic care.
Long-term breast cancer survivors often continue to experience physical and psychological sequelae that affect their quality of life and self-efficacy. Mobile health interventions offer a promising approach to providing personalized and continuous support; however, few have been specifically developed for the long-term survivorship stage. This study describes the design, development, and usability assessment of CUMACA-M, a mobile application aimed at improving quality of life and self-efficacy among long-term breast cancer survivors by supporting the management of persistent sequelae. The research followed the Medical Research Council framework for complex interventions. In the first phase, CUMACA-M was created using a user-centered design approach. Its content was validated by a multidisciplinary panel of experts, whose feedback was complemented by insights from a focus group of long-term breast cancer survivors. In the second phase, usability was evaluated through a mixed convergent design combining the System Usability Scale (SUS), the think-aloud method, and semistructured interviews. Sixteen women recruited from a patient association participated in this stage. The application achieved an average SUS score of 84.84, indicating excellent usability. Participants emphasized the app’s ease of use, personalized content, and intuitive interface, while also suggesting improvements such as clearer tutorials and enhanced visualization. Qualitative findings confirmed the app’s positive influence on organization and motivation for self-care. Overall, CUMACA-M represents a valuable advancement in the digital support of long-term breast cancer survivorship. Its user-centered development and validation by both experts and end users reinforce its relevance and acceptability, while also guiding future enhancements to broaden its functionality and accessibility.
Objectives Population ageing is driving up the prevalence of dementia and intensifying pressure on family carers and long-term care systems. Digital assistive technologies are widely promoted as part of the response, yet they seldom come with a workable reimbursement or implementation model. Using GRACE, an embodied voice assistant for people with dementia (PWD), as an empirical case, this study examines how stakeholder evidence can be translated into setting-specific reimbursement and implementation pathways within the Swiss healthcare system. Methods We conducted 23 semi-structured interviews with healthcare experts (n = 10), professional care staff (n = 9), insurance sector experts (n = 3) and one informal caregiver. German-language transcripts were analysed using codebook-informed thematic analysis and the Business Model Canvas. A policy-translation step distinguished participant statements, author inferences, and literature or policy input. Reporting follows COREQ. Results Stakeholders perceived potential value for PWD in companionship, cognitive activation, orientation and daily structure, and anticipated relief for caregivers. A pivotal finding was that home and institutional settings were expected to carry different requirements and therefore different payer logics. Financing was identified as the dominant barrier: basic mandatory insurance was seen as slow, costly and biased toward curative care, whereas out-of-pocket purchase, shared institutional ownership and supplementary or hybrid arrangements appeared more feasible in the near term. Caregiver acceptance, staff training and demonstrated effectiveness recurred as anticipated enablers. Conclusions Stakeholder evidence can translate perceived value into setting-specific reimbursement and implementation pathways, offering transferable guidance for assistive technologies whose benefits are expected to accumulate across users and care systems. For GRACE, we propose an exploratory pathway beginning with equitably safeguarded out-of-pocket deployment in home care, with movement towards supplementary, hybrid and, potentially, basic insurance contingent on evidence generation. A product split by care setting could align each variant with the payer most likely to benefit. The model is stakeholder-informed and requires validation with PWD and family caregivers. Public Interest Summary Dementia places growing pressure on families and on care services. Voice assistants may support people with early dementia, but it is often unclear who should pay for them and what evidence is needed. We interviewed 23 Swiss professionals, insurance sector experts and one family caregiver about GRACE, a research voice-assistant prototype. Stakeholders anticipated benefits such as companionship, daily structure and caregiver relief, but they had viewed a video rather than used the system. They expected home and residential-care settings to need different product designs and funding arrangements. We therefore propose an exploratory pathway that links each setting and expected benefit to a plausible payer and evidence requirement. Early self-pay or institutional use would require equity safeguards and cannot justify claims of cost savings. Because people with dementia were not interviewed and only one family caregiver participated, the findings describe stakeholder expectations and must be tested directly with intended users and families.
Objectives To characterise archive-derived observed authorisation-holder diversity and evaluate whether a multivariable archive model adds information beyond direct record-completeness fields. Methods This cross-sectional archive study used 779 unique source Names headings in a marketed phase with a recorded exclusivity-expiry date on or before 1 December 2025. Limited diversity was defined as one or fewer normalised company strings with a positive company-attributed authorisation. Measurement error may operate in either direction. Analyses included company-name sensitivity, missing-indicator and complete-case models, nested model-form selection, conditional and full-refit bootstrap uncertainty, a simple completeness benchmark, high-completeness uncertainty, and a Drugs@FDA source-consistency check using maximum-label and union-of-sponsor rules. Results Limited observed holder diversity affected 163 records (20.9%). Removing punctuation and legal suffixes changed holder counts for 64 records but reclassified no outcome. Nested out-of-fold AUCs were 0.852 for recorded volume alone, 0.878 for a simple completeness model, and 0.879 for the full archive model; the full-versus-completeness difference was 0.001 (95% interval -0.013 to 0.013). In 625 records with available DMF and patent blocks, the full-versus-volume AUC difference was 0.005 (-0.018 to 0.027). In the 470-record high-completeness subset with 23 events, AUCs were 0.871 (0.765 to 0.950) and 0.868 (0.789 to 0.931), with a paired difference of 0.003 (-0.059 to 0.060). Drugs@FDA sponsor-union correlations were 0.795 with all-channel archive holders and 0.718 after excluding the archive US FDA channel. Conclusions The full model adds little beyond direct completeness fields and primarily describes record depth within one commercial archive. Its defensible use is to support data curators and regulatory-intelligence analysts in identifying records for manual reconciliation, not to infer market diversity, patent status, supply vulnerability, shortages, or patient risk. Public Interest Summary Complete and current local authorisation registers remain the first and more informative source. This study concerns a commercial archive and asks which records warrant manual reconciliation because company attribution, documentation blocks, or channel coverage may be incomplete or internally inconsistent. The intended user is a data curator or regulatory-intelligence analyst. A flagged record prompts checking of the current local register, company identity, corporate relationships, product granularity, and source-field completeness, followed by correction, annotation, or exclusion of the archive record. The analysis does not identify real-world manufacturers, physical suppliers, shortages, or patient risk.
Objectives To evaluate Australian integrated care models for older adults with chronic conditions and assess their alignment with the Nine Pillars of Integrated Care framework. Methods A scoping review was conducted across MEDLINE, CINAHL, and the Cochrane Library. Peer-reviewed studies published within the past five years were included if they described integrated care models implemented in Australia for individuals aged ≥65 years, or ≥50 years for Aboriginal and Torres Strait Islander peoples. Screening, full-text review, and data extraction were completed using Covidence. Extracted data included model characteristics, outcomes, and levels of integration. A thematic synthesis was undertaken, and each model was mapped against the Nine Pillars of Integrated Care and categorised by level of integration (micro, meso, macro). Results Eight studies representing five distinct integrated care models were included. Most models demonstrated strong patient-centred approaches at the micro level, with some evidence of meso-level service integration. No studies reported macro-level integration. Alignment was strongest in the domains of Population Health and Local Context and Transparency of Progress, Results and Impact. Weak alignment was observed in Resilient Communities and New Alliances and Aligned Payment Systems. The OPEN ARCH model demonstrated the most comprehensive alignment across multiple domains but did not achieve full alignment across all nine pillars. Conclusions Within the limited peer-reviewed evidence identified, integrated care models were reported primarily at the local service and patient-care levels, with little evidence of system-level integration. These preliminary findings identify governance, funding alignment, and cross-sector partnerships as priorities for further research and policy evaluation.
Objectives The governance mechanisms through which public hospitals translate digital transformation into digital health innovation remain underexplored. This study investigates the association between human-centered auditing and digital health innovation in Chinese public hospitals and examines whether organizational ethical culture moderates this relationship. Methods Drawing on perspectives from digital government, digital innovation, socio-technical systems, audit-and-feedback, and organizational ethics, this study conducted a hospital-level cross-sectional survey of 130 Chinese public hospitals between August 2025 and April 2026. Human-centered auditing was operationalized as an entropy-weighted composite index based on Likert-scale indicators encompassing user feedback, ethical and security auditing, interdepartmental coordination, and audit response mechanisms. Regression analyses incorporated organizational and contextual control variables, survey-month fixed effects, robustness checks, and subgroup heterogeneity analyses. Results Human-centered auditing was positively and robustly associated with digital health innovation. Organizational ethical culture did not significantly moderate this association across the sample, indicating that its contextual relevance may depend on broader institutional conditions. Exploratory subgroup estimates varied by hospital level, location, type, and digital health pilot status; however, their directions were not consistently aligned with the pooled interaction estimates, warranting cautious interpretation of subgroup differences. Conclusions This study conceptualizes digital health innovation as an outcome of public-sector digital governance and identifies human-centered auditing as an institutional governance mechanism associated with hospitals’ capacity to integrate user feedback, ethical safeguards, accountability procedures, and digital innovation. Public interest summary Digital technologies are reshaping how public hospitals deliver care, manage services, and interact with patients. This study examined 130 Chinese public hospitals to assess the association between human-centered auditing and digital health innovation. Human-centered auditing encompasses patient and staff feedback, privacy and security reviews, cross-department collaboration, and clearly defined follow-up actions for digital services. Hospitals with stronger human-centered auditing tended to report higher digital health innovation in technology use, management processes, and patient services. Ethical culture did not significantly moderate this association, although ethical values embedded in practical procedures may remain relevant to digital governance within public hospital settings. Overall, digital health outcomes were associated not only with technology investment but also with accountable, user-centered governance arrangements.
Objectives To develop a case-mix adjustment model for inpatient satisfaction scores using the explanatory power method with hospital fixed effects, and to assess the impact of adjustment on hospital performance comparisons. Methods We analyzed cross-sectional survey data from secondary and tertiary hospitals across Gansu Province, China. From 72,656 inpatient responses (68,511 adults after excluding minors and invalid records), the primary analytic sample for all hospital-comparison analyses comprised 38,674 respondents (20,289 self-answer and 18,385 proxy-help) from the 65 hospitals meeting a ≥200-response volume threshold in at least one response mode; the remaining 373 hospitals were excluded because they fell below this threshold. Candidate adjusters were evaluated by predictive power (partial R² from models including hospital fixed effects) and between-hospital heterogeneity, combined into impact factors using the explanatory power method. Coefficient equivalence between the self-answer (n = 20,289) and proxy-help (n = 18,385) subsamples was assessed with a conventional F-test, a hospital-clustered Wald test, and a null-imposed within-hospital permutation test. Inference throughout used hospital-clustered (CR1) variance estimators. Sensitivity analyses compared fixed-effects, random-intercept, and unadjusted specifications; linear versus categorical treatment of ordinal health status; exclusion of the two discharge-measured health-status items; coefficient heterogeneity across proxy-help relationship categories; and a top-box binary outcome. Results Age, general health status (GHS), and mental health status (MHS) were selected as adjusters in both samples; residence was additionally selected for the proxy-help sample, whereas insurance status passed the p-value screen but failed the impact-factor screen. The conventional joint F-test suggested coefficient differences by response mode (p = 0.023), but the cluster-robust Wald (p = 0.374) and permutation (p = 0.162) tests did not detect differences in this dataset. Mean satisfaction nonetheless differed modestly by mode (proxy-help 0.018 points lower than self-answer with the same case-mix; p = 0.004), and a pooled model including a response-mode main effect produced adjusted hospital scores within 0.121 points and 41 of 65 rank positions of the stratified models. Within the proxy-help group, adjustment coefficients were heterogeneous across the assisting members’ relationship categories (p < 0.001), driven by parents-assisted responses. After adjustment, hospital satisfaction scores changed by up to −4.15% and +4.91%, with rank shifts of up to 47 of 65 positions (Kendall’s τ = 0.71). Case-mix adjustment roughly halved funnel-plot overdispersion (φ 19.5 vs 8.94). Fixed-effects and random-intercept coefficients agreed to within 5% in this dataset. The linear treatment of GHS and MHS retained 91.7%–97.1% of the categorical models’ explained variance and left every hospital-level conclusion unchanged. Conclusions Age, GHS, MHS, and—among proxy-help respondents—residence are the key case-mix adjusters for inpatient satisfaction in this provincial context. Robust tests did not detect slope differences between response modes in this dataset, although mean satisfaction differed modestly by mode and the proxy-help group was internally heterogeneous; a pooled model with a response-mode main effect is therefore a reasonable parsimonious choice here, and stratification should remain test-based rather than routine. Case-mix adjustment reduced between-hospital overdispersion in satisfaction scores by roughly half under the specified model, supporting its implementation in provincial performance-comparison programs. These findings derive from the 65 hospitals (n = 38,674) meeting the ≥200-response volume threshold and may not generalize to smaller or lower-volume hospitals excluded by this criterion.
Objectives : Health technology management, particularly under service and rental contracts, represents a challenge for healthcare sustainability. This study describes a Hospital-Based Health Technology Assessment (HB-HTA) data-driven methodology to optimize contract utilization, aiming to support technological consolidation and cost-effectiveness. Methods : A comprehensive relational model was developed, integrating financial, operational, and clinical data from an institutional database. Quantitative criteria, including consumable incidence and equipment utilization rates, informed four strategic actions: decommissioning, aggregation, usage increase, and strategic acquisition. The framework was applied to a fleet of sequencing platforms at a major Italian teaching hospital (TH). Results : The application of the methodology identified redundant units, generally leading to a reduction in the number of contracted equipment. In the case of genetic sequencing platforms, this enabled a reduction of the inventory by 6 physical platforms (45%), demonstrating operational feasibility through workload simulation, resulted in realized annualized savings of €346,000, representing avoided expenditures previously committed to redundant service fees, fulfilling approximately 31% of the hospital’s efficiency target for service contracts. Conclusions : The described framework provides health administrators with an evidence-based tool for Hospital-Based Health Technology Assessment (HB-HTA). It supports informed decision-making in medical device lifecycle management, with the aim of supporting the consistency of healthcare services while maintaining financial resilience.
The European Medicines Agency (EMA) updated its Environmental Risk Assessment (ERA) guideline in 2024 for the first time since 2006, introducing significant changes to environmental protection requirements for pharmaceutical products. This study evaluates the early implementation of the guideline by Marketing Authorisation Holders (MAHs) and National Competent Authorities (NCAs) in Mutual Recognition and Decentralised Procedures. A mixed-methods approach combining an online survey (n=16) and semi-structured interview with a regulatory non-clinical assessor examined implementation challenges, regulatory consistency, and stakeholder perceptions. Results indicate that while the 2024 guideline is viewed as an improvement reflecting current scientific advancements, significant implementation challenges exist. The requirement that generic medicines submit ERAs has placed additional burdens on generic manufacturers. Key barriers include limited data accessibility, insufficient training, inconsistent interpretations across member states and resource constraints. Originator companies with dedicated environmental teams demonstrate better compliance than smaller generic manufacturers. The findings highlight the need for data sharing mechanisms, standardised training programmes, clearer regulatory guidance, and harmonised interpretation across NCAs to support effective implementation while balancing environmental safeguards with pharmaceutical accessibility and the 3Rs principle (Replacement, Reduction, and Refinement).
Objective Fast Healthcare Interoperability Resources (FHIR) has become a critical enabler of modern digital health policy, underpinning national reforms relating to electronic health records, data exchange, and real-time clinical decision support. However, despite its central role in digital transformation, FHIR implementation often falters because national workforces lack the capabilities required to operationalise interoperability mandates. This paper presents a policy-aligned methodological framework for conducting a national workforce needs assessment to inform interdisciplinary FHIR training programs. The framework is designed to support digital health policy implementation by systematically identifying capability gaps and guiding evidence-based training investment. Methods We propose a two-step needs assessment framework integrating general and targeted assessments. The general assessment identifies broad training needs through qualitative interviews, surveys, and market scans, while the targeted assessment refines course offerings based on participant demand, workforce capacity, and contextual factors. A mixed-methods approach combines qualitative thematic analysis and descriptive statistics to triangulate findings. A structured five-component decision-making framework is presented to support transparent training prioritisation aligned with national digital health strategy. Key features This framework is designed to support evidence-based, adaptable training programs aligned with national digital health priorities. This framework addresses identified gaps in existing methodological approaches and is intended to offer advantages over single-phase assessment methods. Conclusion This paper presents a theory-informed methodological framework for designing FHIR training programs, aimed at addressing evolving workforce needs and supporting digital health interoperability globally.
Objectives: To assess the fitness for purpose of reimbursement-oriented and care-oriented electronic health record (EHR) data, identifying which type or combination is most appropriate for informing health policy decisions. Methods: This detailed case study used reimbursement-oriented and care-oriented EHR data to select a cohort of patients treated for colon carcinoma at Jeroen Bosch Hospital in the Netherlands in 2019 and assessed the possibility to determine the date of diagnosis of these patients. Patients were selected through an iterative, multidisciplinary process using structured and unstructured EHR data, followed by manual validation. Multiple inclusion-exclusion strategies were tested, with patient counts and precision and recall percentages reported for each. Results: The combination of structured and unstructured data initially resulted in a patient group of 548 patients. Various combinations of EHR data types and exclusion filters resulted in reduced, but different, numbers of patients. After manual validation, a patient group of 177 patients remained. Assessment of both structured and unstructured data resulted in the determination of the date of diagnosis for all 177 validated patients. Precision and recall were calculated for eight different strategies, resulting in a wide range of precision and recall percentages. Conclusions: Reimbursement-related EHR data offered restricted accuracy for identifying colon carcinoma patients and the start of their care trajectory, but adding clinical, unstructured EHR data improved results. This, however, increases cost, effort, and privacy risks and requires careful balancing. Policymakers should be aware of these limitations and carefully interpret results when using routine health data for secondary purposes.
Objectives: Online mental health communities lack systematic methods to translate discourse into actionable service intelligence. This study operationalizes a Demand-Risk-Response (DRR) framework, applying population health management and health technology assessment (HTA) principles to digital platforms. Methods: Using a dual-cohort design, we analyzed 101,559 posts from a Chinese psychological platform (2020 snapshot) and a post-pandemic replication cohort (2023-2025, n = 29,961). Text-based indices proxied distress severity (Psychological Distress Index, PDI) and response resources (Response Adequacy Index, RAI). Gap analysis identified structural mismatches between high distress and low engagement. Results: Demand concentrated in Personal Growth, Therapy, and Romance. In the 2020 cohort, 27.1% of posts were high-risk, led by Therapy (mean PDI = 1.58) and Self-Improvement (55%). Response adequacy varied significantly by topic (Trending Topics beta = 1.18; Psychoeducation beta =-0.07). Structural gaps affected 9.2% of posts, yielding stable priority rankings (Kendall's tau = 0.94). The 2023-2025 cohort demonstrated lower high-risk prevalence (11.7%) but a persistent gap structure (3.7%). Conclusions: The DRR framework serves as a robust, longitudinal HTA tool. Evolving mismatches demonstrate that digital platform oversight must shift from passive observation to active governance, guiding need-responsive algorithmic routing, interface design, and evidence-based health policy.
Background and objectives The health data space aims to facilitate cross-domain sharing and secondary use of medical data. This study conducts a systematic literature review to identify the structure of construction dilemmas facing the health data space and to synthesize the trustworthy pathways proposed in the existing research. Methods Literature searches were conducted in six databases: PubMed, Web of Science, Scopus, IEEE Xplore, ACM Digital Library and Embase. Following three rounds of screening, 96 studies were included in the final sample and analysed using a narrative synthesis approach. Results The construction dilemmas of the health data space exhibit a linked three-layer structure spanning the individual, legal, and technical levels. In response, the reviewed literature proposes a three-stage trustworthy pathway centered on public trust: trust building through ex-post privacy protection, dynamic authorization, and data format standardization; trust maintenance through property rights control, decentralized architecture, and iterative technical safeguards; and trust enhancement through public value return, trust evaluation, and talent development. Conclusion The dilemmas of the health data space are systemic and interrelated, and the practical feasibility of trustworthy pathways depends on implementation resources, distributional arrangements, and Member State readiness. Future research should evaluate how these conditions shape the effects and equity implications of health data space governance.
