Maintenance Notice

Due to necessary scheduled maintenance, the JMIR Publications website will be unavailable from Wednesday, July 01, 2020 at 8:00 PM to 10:00 PM EST. We apologize in advance for any inconvenience this may cause you.

Who will be affected?

The leading peer-reviewed journal for digital medicine and health and health care in the internet age. 

Latest Submissions Open for Peer Review

JMIR has been a leader in applying openness, participation, collaboration and other "2.0" ideas to scholarly publishing, and since December 2009 offers open peer review articles, allowing JMIR users to sign themselves up as peer reviewers for specific articles currently considered by the Journal (in addition to author- and editor-selected reviewers).

For a complete list of all submissions across all JMIR journals as well as partner journals, see JMIR Preprints

Note that this is a not a complete list of submissions as authors can opt-out. The list below shows recently submitted articles where submitting authors have not opted-out of open peer-review and where the editor has not made a decision yet. (Note that this feature is for reviewing specific articles - if you just want to sign up as reviewer (and wait for the editor to contact you if articles match your interests), please sign up as reviewer using your profile).

To assign yourself to an article as reviewer, you must have a user account on this site (if you don't have one, register for a free account here) and be logged in (please verify that your email address in your profile is correct).

Add yourself as a peer reviewer to any article by clicking the '+Peer-review Me!+' link under each article. Full instructions on how to complete your review will be sent to you via email shortly after. Do not sign up as peer-reviewer if you have any conflicts of interest (note that we will treat any attempts by authors to sign up as reviewer under a false identity as scientific misconduct and reserve the right to promptly reject the article and inform the host institution).

The standard turnaround time for reviews is currently 2 weeks, and the general aim is to give constructive feedback to the authors and/or to prevent publication of uninteresting or fatally flawed articles. Reviewers will be acknowledged by name if the article is published, but remain anonymous if the article is declined.

The abstracts on this page are unpublished studies - please do not cite them (yet). If you wish to cite them/wish to see them published, write your opinion in the form of a peer-review!

Tip: Include the RSS feed of the JMIR submissions on this page on your homepage, blog, or desktop RSS reader to stay informed about current submissions!

JMIR Submissions under Open Peer Review

↑ Grab this Headline Animator

If you follow us on Twitter, we will also announce new submissions under open peer-review there.

Titles/Abstracts of Articles Currently Open for Review:

  • Restrictive Family Relationships as a Mediator of Adolescent Social Media Use Disorder

    Date Submitted: Apr 2, 2026
    Open Peer Review Period: Apr 3, 2026 - May 29, 2026

    We investigated the relationships between academic stress, family relationships, and social media use disorder (SMUD) among middle school students through the ecological systems perspective. We used mixed-methods sequential explanatory. In addition, structural equation modeling showed that academic stress had a significant direct impact on SMUD. Restrictive family relationships significantly mediated this relationship. The proposed model explained 42.6% of the variance in SMUD. Qualitative findings highlighted family patterns and coping and revealed how rigid parental control and limited emotional support reinforced restrictive family dynamics, which in turn further linked academic stress to SMUD. Present research extends Ecological Systems Theory to online behavior and establishes restrictive family relationships as key mediating mechanisms in SMUD development. These findings underscore the importance of family-based interventions that promote open communication and adaptive stress management strategies to mitigate the risk of SMUD.

  • Background: Artificial intelligence-based clinical decision support systems (AI-CDSS) have the potential to help primary care physicians provide higher quality care with rising case complexity and resource constraints. However, evidence on their effectiveness and safety in real-world primary care practice is lacking. Objective: To synthesize evidence on the real-world effectiveness and safety of AI-CDSS compared with usual care or non–AI-based systems in primary care settings. Methods: We conducted a systematic review following the Cochrane Handbook for Systematic Reviews of Interventions and reported the results according to PRISMA 2020. An information specialist designed search strategies and systematically reviewed Medline, Embase, CINAHL, Web of Science, and CENTRAL. Reviewers independently performed the selection and extraction processes. Data extraction was informed by the DECIDE-AI and CONSORT-AI reporting guidelines, and the APPRAISE-AI tool. We independently assessed risk of bias using RoB-2 or ROBINS-I. Due to heterogeneity, we conducted a narrative synthesis across clinician-, patient-, and system-level outcomes as well as safety information. Results: We identified 4085 records and selected 14 records on 11 distinct studies. We identified 11 AI-CDSS across four experimental (high risk of bias) and seven quasi-experimental studies (serious to critical risk of bias). All studies were conducted in high-income countries. Follow-up ranged from one to 12 months. Machine-learning and deep-learning approaches were the most common (n=8/11). The interventions fulfilled different decision-support functions, most frequently treatment or order facilitation (n=8/11). Clinician-level outcomes were most frequently reported (n=8/11), followed by system-level (n=6/11) and patient-level outcomes (n=5/11). Clinician-level findings were mixed, with several studies reporting reductions in administrative or alert burden. System-level outcomes were heterogeneous and sparsely reported, with isolated studies suggesting improvements in cost-effectiveness, service utilization, patient engagement, program completion, and care delivery efficiency. Patient-level effects were limited or inconsistent, with improvements mainly observed in well-defined conditions with established diagnostic and treatment pathways. Safety evaluation was rare (n=2/11) and limited to technical malfunctions or self-reported side effects. Conclusions: AI-CDSS implemented in primary care demonstrate limited and inconsistent effectiveness at clinician-, patient-, and system-level. Safety evaluation relied mainly on self-reported side effects and technical malfunctions, without structured monitoring for AI-related risks. These findings highlight a gap between the pace of innovation and the readiness of healthcare systems for reliable clinical use. Closing this gap will require workflow-integrated system design, timely and standardized evaluation, and active safety monitoring to support trustworthy implementation in primary care.

  • MedEvalArena: A Self-Generated, Peer-Judged Benchmark for Medical Reasoning

    Date Submitted: Apr 1, 2026
    Open Peer Review Period: Apr 1, 2026 - May 27, 2026

    Background: Large Language Models (LLMs) demonstrate strong performance at medical specialty board multiple choice question (MCQ) answering, however, underperform in more complex medical reasoning scenarios. This gap indicates a need for improving both LLM medical reasoning and evaluation paradigms. Objective: To develop an automated framework to evaluate LLM capabilities in medical reasoning. Methods: MedEvalArena is an automated framework in which LLMs engage in a symmetric round-robin format: each model generates challenging board-style medical MCQs, then serves in an ensemble LLM-as-judge bench to adjudicate validity of generated questions, and finally completes the validated exam as an examinee. We compared performance of leading LLMs across the OpenAI, Grok, Gemini, Claude, Kimi, and DeepSeek families on both question generation validity and exam taking performance. Results: Across frontier models, we observe no statistically significant differences in exam-taking performance with mean accuracies 85.7-91.7%, suggesting convergence in medical reasoning ability across frontier LLMs for question-answering tasks. LLM accuracy was comparable to mean human physician accuracy of 85.6% (95% CI: 79.4%-91.7%) and the differences were not statistically significant. We found significant differences between models in question validity rate, with higher question validity rates in questions generated by OpenAI, Gemini, and Claude frontier models (83.3%-94.8%), compared to Kimi, Grok, and DeepSeek models (46.0-63.8%). When jointly considering accuracy and inference cost, multiple frontier models lie on the Pareto frontier with no single model dominating across both dimensions. Conclusions: MedEvalArena provides an automated framework for benchmarking LLM medical reasoning, identifying valid question generation as a more discriminative task compared to question answering.

  • Background: Hospitalized children frequently experience pain and distress. Pain is a multidimensional experience involving both sensory and emotional components, necessitating multimodal management strategies. Socially assistive robots (SARs) have shown promise as non-pharmacological interventions in pediatric care. However, the interaction mechanisms through which SARs influence pain and emotional responses, particularly positive emotion and real-time emotional dynamics during child–robot interaction, remain underexplored. Objective: This study, titled the HAPPY (Hospitalized Assistance for Pediatric Pain Yields) study, aimed to evaluate the association between a SAR-based intervention and postoperative pain in hospitalized children and to examine whether different levels of engagement are associated with changes in real-time emotional dynamics. Methods: A single-group pretest–posttest design was conducted with 37 hospitalized children (mean age 7.35, SD 2.06 years) following tonsillectomy or adenoidectomy. The intervention was structured into three sequential phases: Phase 1 (warm-up/limited engagement), Phase 2 (educational video/passive engagement), and Phase 3 (social interaction/active engagement). Pain was assessed using the Wong-Baker FACES pain scale and observed behavioral FLACC scales. Emotional response, as valence, was measured using an automated facial expression recognition system (FaceReader 10). Changes in pain were analyzed using Wilcoxon signed-rank tests, and differences in emotional valence across phases were examined using the Friedman test with post hoc pairwise comparisons. Results: Self-reported pain significantly decreased from a median of 6 (IQR 4–6) to 4 (IQR 2–4) (P<.001), and observer-rated behavioral pain decreased from a median of 3 (IQR 2–4) to 1 (IQR 1–2) (P<.001). Overall differences in emotional valence across phases did not reach statistical significance (P=.053; Kendall’s W=0.084). However, the V-shaped trajectory of emotional valence was observed, with the lowest values during the passive engagement phase 2 (mean –0.24, SD 0.20) and relatively higher values during the active engagement phase 3 (mean –0.15, SD 0.13). The exploratory post hoc analyses indicated a significant increase in emotional valence from Phase 2 to Phase 3 (adjusted P=.012). Conclusions: SAR-based interventions were associated with reductions in postoperative pain in hospitalized children. Although overall emotional differences across phases were not statistically significant, the observed pattern suggests that active engagement may be associated with more positive emotional responses compared to passive engagement. These findings highlight the potential importance of interaction quality in SAR interventions and provide insight into the processes underlying their clinical effects in pediatric care. Clinical Trial: No

  • Background: Youth mental health needs are substantial: globally, about 1 in 7 adolescents (10–19 years) experience a mental disorder, and suicide is the third leading cause of death among people aged 15–29. Despite need, many young people delay or avoid care due to stigma, privacy concerns, cost, and difficulty accessing timely services. Chat-based support can lower the threshold for disclosure, but safety-critical statements require transparent risk stratification and clear pathways for clinician oversight. Language further affects safety in multilingual communities: Cantonese is comparatively under-resourced for natural language processing, and its colloquial orthography, idioms, and pervasive code-switching can distort risk-cue interpretation when models are trained primarily on English and standard written Chinese. Objective: We developed EmoBot, an explainable three-tier (Tier 1–3) risk stratification pipeline and nurse-/counsellor-facing dashboard for Cantonese-speaking youth, and evaluated agreement with expert triage labels and boundary error patterns. Methods: EmoBot uses a hybrid framework combining semantic exemplar retrieval and constrained large language model (LLM) reasoning. An expert-authored, de-identified Cantonese reference corpus is embedded with Sentence-BERT and indexed for nearest-neighbor retrieval (top-k=5). In parallel, the DeepSeek API chat model (deepseek-chat; DeepSeek-V3.2) generates a structured record (tier, category, cues, rationale) from the user message, tier/category definitions, and retrieved exemplars as in-context annotated examples. Inference settings were fixed (temperature=0.1; presence_penalty=0; max_tokens capped) and no fine-tuning was performed. If retrieval and LLM disagree, EmoBot outputs the higher tier while displaying provenance-linked evidence in the dashboard to support nurse-in-the-loop verification and escalation. Two sequential validation sets were used (set 1: n=50 pilot stress-test; set 2: n=336 expanded validation; total n=386). Three mental-health–trained experts rated each statement blinded to model output; majority consensus (≥2/3) defined the reference label. Results: Human majority consensus was achieved for 382/386 statements (99.0%). On consensus-labeled items (n=382), EmoBot matched expert tiers with 95.8% accuracy (366/382; 95% CI 93.7–97.6) and 95.2% macro F1 (95% CI 92.7–97.4). Performance increased from set 1 (80.9% accuracy) to set 2 (97.9% accuracy). All misclassifications were adjacent-tier (±1) with no Tier 1↔Tier 3 confusions; Tier 3 detection remained high (F1 94.7% and 98.6%). Conclusions: Explainable, conservative hybrid triage with provenance-linked evidence can closely align with expert judgment for Cantonese youth help-seeking text while supporting nurse-in-the-loop review and escalation. The architecture (localized exemplar corpus + structured model outputs + auditable dashboard) is adaptable to other low-resource languages and multilingual settings where safe risk triage requires both linguistic localization and clinical accountability. Clinical Trial: Not applicable (nonrandomized expert validation study; no clinical trial registration required).

  • Public Perception of Health Care Before, During, and After COVID-19: A Longitudinal Analysis of Online Reviews

    Date Submitted: Apr 1, 2026
    Open Peer Review Period: Apr 1, 2026 - May 27, 2026

    Background: Online reviews of health care services represent a growing source of unsolicited, citizen-generated data that can complement traditional instruments for monitoring public perception of health systems. However, longitudinal analyses examining how citizen perception evolved before, during, and after the COVID-19 pandemic remain scarce, and existing studies have rarely differentiated between levels of care. Objective: This study aimed to examine the longitudinal evolution of public perception of a regional public health system over a ten-year period, with particular attention to differences between primary care and hospital services, and to assess whether the COVID-19 pandemic produced a temporary disruption or a more persistent structural shift in citizen evaluation of health care. Methods: A retrospective longitudinal observational study was conducted using 47,589 online reviews of 812 public health care facilities in Andalusia, Spain, collected from Google Maps and covering the period 2016–2025. Reviews were classified as positive or negative based on star ratings, validated against manual annotation using Cohen's kappa. The proportion of negative reviews was analyzed across three periods: pre-pandemic (2016–2019), pandemic (2020–2021), and post-pandemic (2022–2025). Structural breaks were identified using change-point detection analysis. Logistic regression models with robust standard errors clustered at the facility level were used to quantify differences in negative sentiment across levels of care and over time. Results: The proportion of negative reviews increased from 38.7% in the pre-pandemic period to 73.7% during the pandemic, remaining elevated at 66.5% in the post-pandemic period. Change-point detection identified March 2020 as a major structural break. The pandemic had markedly different effects across levels of care: negative reviews in primary care rose from 34.8% to 81.9% during the pandemic, remaining at 75.7% post-pandemic, whereas hospital care showed a more moderate increase from 43.8% to 55.6%, remaining stable thereafter. Logistic regression models confirmed that the trajectory of negative perception in primary care diverged significantly from hospital care during and after the pandemic, with interaction terms indicating substantially higher odds of negative reviews in primary care during the pandemic (OR = 5.28, 95% CI 3.95–7.07) and post-pandemic periods (OR = 3.66, 95% CI 2.66–5.03). Conclusions: The findings indicate that the COVID-19 pandemic was associated with a persistent structural shift in public perception of health care services rather than a temporary fluctuation, and that this shift was disproportionately concentrated in primary care. The sustained deterioration in citizen perception of primary care observed years after the acute crisis suggests that post-pandemic recovery strategies should explicitly address the post-crisis phase and prioritize the relational and communicative dimensions of primary care alongside structural capacity. Large-scale digital trace data offer a scalable and continuous complement to traditional patient satisfaction instruments for monitoring health system legitimacy over time.

  • Externalized Living Memory: Structuring Clinical Knowledge for the Age of AI Agents

    Date Submitted: Apr 1, 2026
    Open Peer Review Period: Apr 1, 2026 - May 27, 2026

    As AI agents become increasingly capable of autonomous action in health care, a prerequisite remains underaddressed: the persistent, structured memory that makes such action contextually meaningful. Clinicians face cognitive overload not from any single task but from the erosion of decision context over time. Existing tools—personal knowledge management frameworks, LLM built-in memory, and autonomous agents—each address parts of this problem but leave gaps in auditability, portability, or contextual persistence. This Viewpoint argues that memory should precede action: before AI agents can act meaningfully, they need persistent, human-controlled context. We describe externalized living memory—a structured knowledge base that both human and AI can read and write—as it emerged from the first author's practice as a cardiovascular radiologist and division chief. The approach is organized as a layered architecture with a routing table for scalable context loading and a governance hierarchy for sustainable maintenance. We illustrate the approach through clinical vignettes, compare it with existing solutions, and discuss limitations including the small-team evidence base and maintenance costs. An open-source implementation with templates and setup instructions accompanies this paper.

  • Background: The dissemination of accurate, timely pharmaceutical product information is foundational to patient safety. Electronic labeling (e-labeling) has emerged as promising digital health approach that enables real-time updates, enhanced accessibility, and integration within digital health information ecosystems. Despite growing global adaptation, regulatory frameworks remain fragmented across jurisdictions, creating disparities in health information equity and posing challenges to patient safety in interconnected healthcare environments. Objective: This study aimed to systematically identify and compare national pharmaceutical e-labeling policies across major regulatory jurisdictions, assess regulatory readiness and policy maturity, characterize implementation strategies, identify critical governance gaps, and provide evidence-based recommendations for a harmonized global framework advancing digital health equity. Methods: We systematically searched and reviewed publicly available pharmaceutical e-labeling policy documents from International Coalition of Medicines Regulatory Authorities (ICMRA) member countries covering January 2012 to January 2024. Two independent reviewers screened and classified documents as “legal” (binding regulations) or “non-legal” (guidance) based on regulatory authority. A two-step analysis was employed: a hybrid thematic coding approach combining deductive coding based on three predetermined domains (scope and implementation, operational infrastructure, advanced digital health considerations) with inductive coding to generate nine themes from the data, followed by a comparative analysis of domain coverage, regulatory level, and regional distribution. Inter-rater reliability was assessed using Cohen's kappa coefficient. Policy maturity was classified across five levels based on regulatory authority and scope. Results: Among 22 ICMRA member countries, only 11 (50.0%) had accessible e-labeling policies, yielding 34 documents: 5 legal (14.7%), 29 non-legal (85.3%). All 11 countries addressed foundational implementation, scope and operational infrastructure, but only 9 (81.8%) covered advanced considerations including accessibility, interoperability, with electronic health records, or personalized information. Legally binding framework for basic implementation existed in merely 4 countries (36.4%) and only 2 countries (18.2%) had binding provisions for advanced features. The most sophisticated digital health capabilities, interoperability with clinical systems and personalized patient information, were addressed in only 3 countries (27.3%), through non-legal guidance. A sequential policy progression trend was observed, with nations prioritizing operational infrastructure before advancing to more complex considerations. Dual independent coding achieved excellent reliability (Cohen's kappa ≥0.95 for all domains). Conclusions: Global pharmaceutical e-labeling policy development remains substantially fragmented with critical gaps in legal binding frameworks, technical standards, and patient-centered digital health integration. Only one-third of major regulatory markets possess legally binding implementation frameworks, while advanced interoperability and personalization capabilities remains largely unexplored. These findings highlight the urgent need for international harmonization through inter-agency collaboration, including adoption of binding minimum standards, machine-readable format mandates, and interoperability protocols aligned with global health informatics standards.

  • Background: Digital tools are increasingly used to organize, analyze and report behavioral science data, informing interventions addressing health, sustainability and other global challenges. As these tools proliferate, there is a need for methods that evaluate their usability, acceptability and influences on uptake, using theory-informed behavioral approaches. Objective: This study aims to explore: (1) the usability and acceptability of the Theory and Techniques Tool (TaTT; tool for intervention development), alongside the Behaviour Change Intervention Ontology (BCIO; a classification framework for behavior change interventions) online tools, and (2) barriers and enablers to using them to specify behavior change techniques (BCTs) and their delivery, using a behavioral model. Methods: Fourteen intervention development experts participated in a think-aloud task using the TaTT and BCIO’s tools to identify BCTs and intervention delivery in a voter-related intervention development scenario. This was followed by interviews to identify barriers and enablers applying the tools, structured around the Capability-Opportunity-Motivation Behaviour (COM-B) Model. Transcripts were analyzed using inductive thematic analysis, and identified barriers and enablers were mapped to COM-B constructs. Results: We identified four usability and acceptability themes: “Information and presentation clarity” (e.g., unclear condensed information); “Navigational ease within and between tools” (e.g., difficulty moving between pages); and “Poor site performance of BCIO tools” (e.g., site crashes) and “Utility of the tools” (e.g., helpful search functions). Four COM-B constructs were identified as barriers and enablers: Psychological capability (e.g., limited knowledge of ontology structure), Physical opportunity (e.g., lack of real-time guidance), Social opportunity (e.g., mixed views on BCIO acceptance) and Reflective motivation (e.g., perceived value of tool). Conclusions: The findings highlight where TaTT and BCIO tools can be improved and where additional guidance is needed. The study also demonstrates a method for exploring usability that can be applied to other digital tools in behavioral sciences. Clinical Trial: N/A

  • Background: There is a considerable risk of stigmatization and harm when sensitive topics such as psychiatric diagnoses, substance abuse, and self-harm are recorded in electronic health records, especially since federal laws like the 21st Century Cures Act now require patients to have access to their own clinical notes. Discharge summaries are particularly problematic since they combine all hospital experiences and perform concurrent administrative, legal, and patient-facing tasks, although they are still not well researched in sensitive topic studies. There is a major methodological vacuum in understanding how the documentation of certain sensitive topics changes over time because current NLP techniques have concentrated on single-topic detection tasks with little attention to prevalence measurement or longitudinal documentation patterns. Objective: To evaluate an NLP based framework for the automated detection of a predefined set of sensitive topics and measuring their change in prevalence over time using the MIMIC-IV database. Methods: Discharge summaries from the MIMIC-IV database were filtered using ICD codes to identify psychiatrically relevant admissions, resulting in 2670 notes from 2108 distinct patients. Both explicit keyword-based and implicit semantic mentions of each sensitive subject category were found using a dual NLP detection framework. Normalized mention counts were used to compute weighted prevalence, which was then examined at the note and patient levels. Fisher’s exact test, McNemar’s exact test, and Benjamin Hochberg FDR correction were used to evaluate temporal change. Results: The weighted prevalence varied across sensitive categories, ranging from 2.68% to 18.93%. Patient level prevalence was consistently higher than note level prevalence across all categories. No statistical significance was observed after FDR correction, suggesting stable documentation over the study period. Improving, worsening, and absent trajectories remained relatively steady independent of overall category predominance, but persistent and mixed documentation patterns showed the most variance among categories, according to trajectory analysis. Conclusions: This study demonstrates the feasibility of a scalable NLP framework for multi-topic sensitive topic detection and prevalence tracking in clinical discharge summaries. The findings highlight the value of combining note-level and patient-level analyzes and provide a foundation for future work examining the impact of policy changes on sensitive topic documentation practices.

  • Background: Generative artificial intelligence (GenAI) has increasingly entered psychiatric practice through patient-facing chatbots, self-help tools, and clinician-facing workflow support. Although prior research has examined clinicians’ attitudes, readiness, and anticipated use cases, less is known about how frontline encounters with GenAI shape psychiatrists’ interpretations and implementation priorities. Healthcare foresight also remains methodologically underdeveloped and has focused mainly on external signals, overlooking clinically consequential signals emerging from everyday practice. This gap is especially important in psychiatry, where GenAI-related benefits and harms may depend on patient vulnerability, crisis sensitivity, and the therapeutic relationship. Objective: To qualitatively examine how South Korean psychiatrists described clinical experiences with GenAI, how they interpreted its roles and limits in psychiatric care, and what implementation priorities they emphasized. Selected concepts from horizon-scanning informed the organization of the analysis by orienting attention to practice-based signals, interpretive patterns, and implementation priorities. Methods: In this qualitative descriptive study, directed content analysis and codebook-based thematic synthesis were used to analyze responses to 3 open-ended survey questions administered to members of the Korean Neuropsychiatric Association. Invitations were distributed through the association’s official email system from October 27 to December 26, 2025. The qualitative analysis included respondents who provided an interpretable response to at least 1 item. The questions addressed (1) GenAI-related clinical experiences, (2) perceived advantages and limitations of chatbot-based AI relative to human therapists, and (3) priorities for the safe introduction of GenAI into mental health care. An exploratory participant-level cross-question thematic alignment analysis was also conducted to examine recurring adjacent-item pairings across the experience-interpretation-priority sequence. Results: Of 408 total survey respondents, 311 respondents provided a meaningful response to at least 1 open-ended item. Psychiatrists described GenAI as a clinically ambivalent technology whose implications depended on context, intensity of use, and patient vulnerability. Practice-based signals clustered around patient-led use, clinician-led use, GenAI as a relational object, and GenAI-mediated changes in the patient-clinician interface, with high-risk and destabilizing scenarios cutting across these domains. Experiences ranged from self-help, emotional reflection, triage, and workflow support to overreliance, conflict with clinical authority, reinforcement of distorted or delusion-like beliefs, and suicide- or self-harm-related risk. Respondents viewed GenAI as potentially useful as an adjunct, but also as relationally limited and unacceptable as a replacement for human therapists. Implementation priorities centered on governance, crisis and vulnerability safeguards, technical reliability and clinical validation, and education, supervision, and structural readiness. Cross-question analysis suggested recurrent alignments between frontline signals, a view of GenAI as standardized and tireless but relationally thin, and governance- and validation-oriented implementation priorities. Conclusions: In this qualitative descriptive study, GenAI emerged in psychiatric practice as an access tool, a workflow aid, and, at times, a competing interpretive reference point in clinical encounters. The key implementation challenge is therefore not whether psychiatry will encounter GenAI, but how its use should be bounded, supervised, and governed in light of patient vulnerability, psychiatric risk, and the relational demands of care. Clinical Trial: Clinical Research Information Service (CRIS) KCT0011712; https://cris.nih.go.kr/cris/search/detailSearch.do?seq=32747&search_page=M&search_lang=E&class_yn=

  • Do androids dream of lived experience? A call for human connection in collaborative research amidst the growth of AI

    Date Submitted: Mar 30, 2026
    Open Peer Review Period: Mar 31, 2026 - May 26, 2026

    Professionals, leaders, and institutions in healthcare and health research are rapidly adopting and integrating AI systems and chatbots into their regular work, but this poses risks for patients in the case of patient and public involvement and engagement (PPIE). AI offers economical solutions for overstretched health systems and burned-out staff, already shows strengths in speeding up more long-term and minute research practices, and providing unique accessibility accommodations. However, AI can also be used to create personas and virtual PPIE panels, which can speak completely or partially for human patients with lived experience of conditions, thus minimising, distorting, or erasing their voices from collaborative research processes. AI pose risks through several distorting factors, including hallucinations, overconfidence, sycophancy, bias, sexism, and racism. Staley and Barron have argued that learning is the greatest outcome of PPIE. However, if researchers, professionals, and staff use AI chatbots in conjunction with or in lieu of human collaborators, the amount of learning that takes places is greatly reduced, according to AI expert and cultural critic, Ethan Mollick. In conclusion, we provide a checklist to guide professionals and researchers in ethical and responsible uses of AI that preserves the voices and roles of patients, members of the public, and lived experience.

  • Digital Illness Narratives Among Young Chinese Patients With Diabetes: A Thematic Analysis of Social Media Posts

    Date Submitted: Mar 29, 2026
    Open Peer Review Period: Mar 30, 2026 - May 30, 2026

    Background: Chronic illness often disrupts individuals’ everyday lives and sense of self, particularly among young patients navigating identity formation. In the context of expanding digital media, online platforms have become important spaces where patients articulate illness experiences and seek support, yet how these processes unfold in non-Western contexts remains underexplored. Objective: This study investigates how young Chinese diabetics negotiate their experiences of diabetes, and reconstruct their identities in response to chronic illness in digital spaces. Methods: A thematic analysis was conducted on 303 narrative posts from RedNote, a Chinese social media platform featuring intimate, user-generated storytelling. Methods: A thematic analysis was conducted on 303 narrative posts from RedNote, a Chinese social media platform featuring intimate, user-generated storytelling. Results: Four narrative types were identified: chaos (cognitive dissonance and disordered life), stigma (social withdrawal and discrimination), resilience (emotional fortitude and self-routinization), and solidarity (familial solidarity and reciprocal digital community). Conclusions: These narratives reveal diverse strategies through which young diabetics manage disruption and regain agency. The study extends biographical disruption theory by conceptualizing disruption as a dynamic, relational process of digital re-storying, offering insights for culturally sensitive health communication and online patient support.

  • Use of Digital Devices and Change in Loneliness over Ten Years: Findings from a Swedish Population-based Longitudinal Study on Aging

    Date Submitted: Mar 30, 2026
    Open Peer Review Period: Mar 30, 2026 - May 25, 2026

    Background: Loneliness is recognized as a global health threat. Older adults are vulnerable to loneliness due to life-changes common in old age. While individual risk factors of loneliness in old age are well-documented, contextual factors are scarcely explored, such as digitalization. Rapid digitalization underscores the need to explore the long-term effect of use of digital devices on loneliness. Objective: To explore whether and how the use of digital devices is associated with changes in loneliness over a ten-year period in a population-based sample of older adults. Methods: Data were obtained from the Swedish Adoption/Twin Study of Aging (SATSA) (N=771; mean age 69.4 years). Digital use use and loneliness was assessed across five waves between 2004 and 2014. Age, sex, education, living situation, self-rated health, and the personality trait openness were assessed at baseline. Growth Mixture Modeling was employed to identify latent trajectories of loneliness, and multinomial logistic regression predicted class membership based on baseline digital use and covariates. Results: Three latent loneliness classes were identified: Class 1 (10.5%; high intercept and significant increases in loneliness), Class 2 (33.2%; intermediate stable loneliness), and Class 3 (56.3%; low stable loneliness). Higher digital use at baseline significantly decreased the odds of belonging to the high-increasing loneliness group (Class 1) compared to the low-stable group (Class 3; OR 0.76, p=0.02, CI: 0.60-0.96). When comparing the two groups with higher loneliness levels (Class 1 vs. Class 2), digital use was the only significant predictor; higher use lowered the odds of experiencing increasing loneliness over time (OR 0.77, p=0.04). Differences between classes were not explained by the personality trait of openness to experience. Conclusions: Higher use of digital devices is associated with lower and more stable levels of loneliness over time. These findings suggest that digital technology might serve as an effective non-invasive tool to combat loneliness in older populations.

  • Opioid-related Drug-Drug Interactions and Harm to Hospitalized Patients: A Retrospective Multicenter Cohort Study

    Date Submitted: Mar 29, 2026
    Open Peer Review Period: Mar 30, 2026 - May 25, 2026

    Background Opioid-related drug–drug interactions (DDIs) are common in hospitalized patients and can lead to serious harm, especially when opioids are combined with central nervous system depressants. Electronic medical records (EMRs) often trigger DDI alerts to warn clinicians of potential DDIs, but the effect of DDI alerts on clinically relevant opioid DDIs and related patient harm remains uncertain. This study evaluated whether EMR-integrated opioid DDI alerts reduce clinically relevant interactions and associated harms in routine hospital care. Objective This study aimed to address this gap by determining whether introducing these alerts reduces the prevalence of potentially and clinically relevant opioid-related DDIs, as well as the rate of DDI-related patient harm in hospitalized patients. Methods This retrospective cohort study was a secondary analysis of a multicenter quasi-experimental controlled pre–post evaluation of EMR implementation across five Australian hospitals. Adult inpatients were randomly selected from all patients who stayed in study hospitals for a one-week period six months before and six months after EMR implementation. Inpatients were included if they had at least one prescribed and administered opioid and one concurrent medication. Interruptive opioid DDI alerts were active only at intervention sites post-EMR. Potential DDIs were identified using Stockley’s Interaction Checker; pharmacists adjudicated clinically relevant DDIs, and clinical pharmacologists assessed DDI-related harm and causality. Clustered logistic regression with generalized estimating equations, adjusting for demographic and clinical variables, estimated the effect of alerts involving opioids on three outcomes: clinically relevant opioid DDIs (primary), any potential opioid DDI, and opioid DDI-related harm. Results Of 1,144 patients prescribed an opioid, 847 (74.0%) had at least one potential opioid DDI and 548 (47.9%) had at least one clinically relevant DDI. EMR alerts were associated with no significant change in clinically relevant DDIs (adjusted odds ratio 1.06, 95% CI 0.72–1.55; p=0.75). There was a significant reduction in potential opioid DDIs (adjusted odds ratio 0.55, 95% CI 0.41–0.74; p<0.001). Of all patients, there were 11 patients with a total of 38 DDIs experienced harm (0.6% of potential and 1.1% of clinically relevant DDIs), with most DDIs involving pharmacodynamic interactions with concomitant CNS depressants. Conclusion EMR opioid DDI alerts reduced overall exposure to potential DDIs but did not decrease clinically relevant interactions or related harm. The low rate of harmful events highlights the limited clinical value of current alert systems and the burden of low-value warnings.

  • Digital Health Among Educators: A Scoping Review

    Date Submitted: Mar 30, 2026
    Open Peer Review Period: Mar 30, 2026 - May 25, 2026

    Digital health technologies are increasingly used to support health management, yet research focusing on educators’ engagement with digital health remains limited. Given that educators’ health literacy and digital practices influence both their own wellbeing and their students’ health behaviors, understanding their interaction with digital health tools is essential. This scoping review maps the current literature on educators’ digital health and identifies gaps in knowledge. Following the Arksey and O’Malley framework, a systematic search for peer-reviewed articles yielded 17 eligible studies. Thematic analysis revealed three themes: 1) educators’ digital health literacy and its correlates; 2) educators’ experiences and challenges in using digital health technologies; and 3) professional support and interventions for educators’ digital health. The findings indicate uneven levels of educators’ digital health literacy, mixed experiences shaped by usability and contextual constraints, and insufficient institutional support for sustained engagement. This review highlights the need for targeted capacity-building efforts and context-sensitive interventions to enhance educators’ digital health competencies and to inform future research and policy development.

  • Operationalizing ethical review protocols for health data access: a scoping review of tools and frameworks

    Date Submitted: Mar 27, 2026
    Open Peer Review Period: Mar 30, 2026 - May 25, 2026

    Background: The European Health Data Space (EHDS) will substantially increase cross-border health data sharing in the EU, tasking Health Data Access Bodies (HDABs) with the legal and ethical assessment of health data access requests. However, ethical evaluation of data access is often criticized as an opaque, inconsistent and difficult to operationalize, relying on list of principles with limited procedural guidance. As data access requests increase, the lack of standardized and actionable ethics review processes risks undermining transparency, consistency, and trust in health data governance. Objective: This scoping review (ScR) aimed to identify, map and synthesize existing tools and frameworks designed to operationalize ethical evaluation of health data access. It focused on tools moving beyond conceptual reflection by providing standardized, repeatable, or quantifiable processes that support actionable decision-making. Methods: The ScR was conducted in accordance with PRISMA-ScR guidelines, searching both academic databases and grey literature. Search results were imported into Covidence for screening when possible. For sources incompatible with Covidence, screening and data extraction were conducted manually, using Excel sheets. Two reviewers independently screened titles, abstracts and full texts, with disagreements resolved by consensus. To minimize bias, reviewers were aware of each other’s involvement but could not see each other’s decisions during the initial screening phase. Data were extracted and thematically analyzed to examine tool characteristics, intended users, evaluative criteria, patient involvement, use of quantification, and real-world applications. Results: The ScR retrieved 2215 results, of which 1887 were unique (82.15%). A total of 13 full text studies were included and analysed based on the following criteria: the type of tool, the actor meant to use it, what it measured, its actionable components, whether it contained quantified or quantifiable components compatible with (partial) automation, and real-life applications. These 13 tools differed in their approaches, varying from multiple choice questionnaires (n=5), qualitative questionnaires (n=3) and decision aids, frameworks or matrixes (n=5). Most of these tools were directed at data users (n=10) and mainly aimed at guiding reflection or generating reports for further ethics assessment (n=7). Very few directly involved the public or patients in their development (n=3). Some provided means to classify data use into risk categories with an associated lower or higher level of ethical scrutiny (n=5), at times incorporating quantification (n=5) in the review process. Finally, most tools had limited or no documented real-world implementation (n=9). Conclusions: An increasing number of tools and frameworks aiming at standardizing and operationalizing ethical assessment in data access are being developed in recent years, however their effectiveness remains untested in most cases. As data flows increase in the EHDS, and consequently, the need for streamlined ethics becomes more apparent, hybrid models incorporating both quantitative and deliberative components may play an important role in tackling the challenges associated with ethical data access management. Clinical Trial: N.A.

  • Barriers and Facilitators to Implementing Artificial Intelligence in Intensive Care Units in China: A CFIR-Based Qualitative Study from Nurse Managers’ Perspectives

    Date Submitted: Mar 26, 2026
    Open Peer Review Period: Mar 27, 2026 - May 22, 2026

    Background: Artificial intelligence (AI) has shown significant potential in ICU nursing practice, enhancing efficiency, making decisions, and patient safety. However, evidence regarding the implementation factors of AI in ICU nursing remains limited, particularly from the perspective of nursing leadership. Objective: This study aimed to explore perceived barriers and facilitators to implementing AI in ICUs in China from the perspectives of ICU nursing managers, guided by the Consolidated Framework for Implementation Research (CFIR). Methods: A qualitative study using semi-structured, face-to-face interviews was conducted with 11 ICU nursing managers from tertiary hospitals across seven geographic regions in China. Interview questions were informed by the CFIR framework. Data were audio-recorded, transcribed verbatim, and analyzed using a combined deductive–inductive approach with NVivo software. CFIR constructs were coded and rated as barriers, facilitators, or neutral factors. Results: A total of 20 factors were identified across five CFIR domains, including 5 barriers, 13 facilitators, and 2 neutral influencing factors. Key barriers included high implementation costs, limited adaptability and complexity of AI systems, ethical and privacy concerns, shortages of interdisciplinary talent, and communication challenges between clinical and technical teams. Major facilitators encompassed perceived relative advantages of AI, supportive national policies, leadership engagement, a positive implementation climate, readiness for implementation, and nurses’ self-efficacy. Conclusions: AI implementation in Chinese ICUs is a complex socio-technical process influenced by multilevel contextual, organizational, and individual factors. While nursing managers hold generally positive attitudes toward AI, addressing structural, ethical, and workforce-related challenges is essential for sustainable integration. These findings provide theory-informed insights to support context-sensitive implementation strategies for AI in critical care nursing practice. Clinical Trial: None.

  • The development of artificial intelligence applications in healthcare has historically required multidisciplinary technical teams, substantial programming expertise, and prolonged development cycles, effectively excluding clinicians, the professionals with the deepest contextual understanding of clinical problems, from the role of active builder. The emergence of no-code development platforms, vibe coding tools, and AI workflow orchestration systems has fundamentally altered this landscape, creating a realistic pathway for clinicians, medical educators, and healthcare researchers to design, build, and deploy custom AI-powered applications without traditional programming knowledge. This tutorial presents a structured five-level clinician-builder development framework spanning conversational chatbot construction using no-code, vibe coding and AI coding assistants. The Clinician-Led Evidence-grounded AI Development framework is proposed as a reproducible methodology for guiding clinician-initiated AI development projects from clinical problem identification through deployment and monitoring. This tutorial demonstrates that the technical barriers separating clinical insight from AI application development have been substantially lowered, and that the tools required for clinician-led innovation in healthcare AI are now accessible, practical, and ready for responsible deployment

  • Background: Background: Artificial intelligence (AI) is quickly becoming a key part of digital health systems in oncology, supporting activities like cancer screening, clinical decision-making, and patient care management. Although AI has the potential to enhance care quality and efficiency, its adoption at cancer centers varies widely, raising concerns about disparities in digital health access and capacity. Objective: Objective: This research investigates the multiple factors influencing AI adoption as part of digital health implementation at National Cancer Institute (NCI)-designated cancer centers across the U.S., focusing on institutional readiness, policy environment, and geographic spread. Methods: Methods: A national dataset of 75 cancer centers was assembled using public sources to track AI use in screening, treatment, and patient care. AI adoption was measured as a composite index (0-3), indicating integration across clinical areas. Spatial patterns were analyzed with Moran’s I, and multilevel ordered logistic regression models examined links between AI adoption, institutional features (like number of physicians, hospital beds, center type), and contextual factors (such as socioeconomic status and state politics). Results: Results: No significant clustering of AI adoption was found geographically, implying limited regional diffusion. The size of the physician workforce was the most consistent predictor of AI adoption, emphasizing that organizational readiness is a key driver. Policy environment also influenced adoption: comprehensive cancer centers in Republican-controlled states showed higher AI uptake. Socioeconomic status at the community level was not significantly related. Conclusions: Conclusions: This study identifies institutional capacity and policy environment as primary constraints on scalable innovative digital health implementation in cancer institutions. These results point to structural barriers to broad digital health deployment and indicate that advancing AI-enabled cancer treatment will need focused investments in institutional capacity and policy support. Without these efforts, disparities in digital health infrastructure could restrict equitable access to AI-driven innovations in oncology.

  • Background: The Military Healthcare System mandates medical documentation at all echelons of care, however care providers in combat situations must prioritize lifesaving measures over record-keeping, leading to information gaps. Effective human machine teaming (HMT) solutions designed to autonomously document care delivery will serve as future force multipliers in tactical combat casualty care (TCCC) environments. Objective: To address this challenge, the United States Army Institute of Surgical Research commenced an effort to prototype HMT systems designed to passively document care delivery in TCCC environments. However, common artificial intelligence (AI) performance evaluation methods do not represent the temporal, repetitive and context dependent nature of real-world TCCC delivery. It was essential to conduct comprehensive assessments of the AI functions in a timely, synchronized manner. During the initial prototyping phase, developers were provided with annotated datasets from seventy-five TCCC simulations and given six months to develop their algorithms. Methods: To assess performance, the research team leveraged a reserved dataset. In the first phase of the assessment a standardized, repeatable performance methodology and framework was leveraged to evaluate individual algorithms that detect: (1) injury location; (2) visible medical objects, and (3) treatments administered by the care provider. Detection effectiveness included four metrics: modified accuracy; precision; recall; and F1 scores. Algorithm processing efficiency was evaluated by calculating lag time scores. A final composite score was used to quantify performance differences among the algorithms within a specific detection category. The second phase of the evaluation integrated multiple algorithms into a centralized orchestration framework; enabling synchronized execution and consolidated outputs. System resource usage and throughput metrics were evaluated to characterize the efficiencies. Quantified memory consumption and processing unit resource utilization were assessed, followed by benchmarking the edge compute orchestration framework. Results: Results to date revealed that Medical Object Detection achieved the highest performance (mean F1≈0.42, range 0–0.71). Injury Detection and Localization showed lower performance (mean F1≈0.27, range 0–0.60), with higher recall than precision. Medical Procedure Detection yielded procedure-level mean F1 scores from 0.00 to 0.31 and simulation-level means from 0.00 to 0.33. Stronger results were observed for Nasopharyngeal Airway and Chest Seal Application (mean F1≈0.28, 0.31) medical procedures. The results to date are preliminary and serve as illustrative examples of the evaluation framework outputs. Conclusions: The preliminary results highlight the evaluation framework’s end-to-end, standardized results across core algorithm functions. While algorithm performance to date is modest, the framework demonstrates its capacity to capture both variability and recurring patterns across simulations, thereby highlighting strengths, limitations, and areas requiring refinement. It enables reproducible, cross data set comparisons, allowing evaluators to quantify algorithm performance. By leveraging both simulation-level evaluations and detection specific performance aggregated across simulations, the framework enables targeted identification of underperforming areas, supporting iterative and strategic AI model refinement.

  • Background: Artificial intelligence (AI) systems designed to enhance polyp detection during colonoscopy have shown promise in clinical trials, but the extent to which various systems improve detection of polyps of different sizes remains unclear. Objective: The objective of this article is to compare efficacy of various AI-assisted systems during colonoscopy. Methods: We searched PubMed, Cochrane Library and Web of Science for randomized controlled trials (RCTs) comparing AI-assisted colonoscopy with white-light colonoscopy, then literature screening was conducted and network meta-analysis (NMA). NMA was conducted using StataMP 14 and Review Manager 5.3 was used for the quality assessment of included studies. Results: NMA of 13 RCTs (4,156 participants) revealed marked size-dependent efficacy of AI-assisted systems for increasing mean polyp detection. For diminutive polyps, EndoScreener ranked first [SUCRA 77.7%, probability of being best 33.3%] and significantly increased mean polyp count versus standard colonoscopy (SMD 1.22, 95%CI 0.08 to 2.36). For small polyps (6-9mm), standard colonoscopy unexpectedly achieved highest ranking (SUCRA 91.5%), with AI systems showing minimal incremental value. For large polyps (≥10mm), CAD EYE and standard colonoscopy performed comparably (SUCRA 78.9% vs 80.0%), with no AI system demonstrating significant improvement. No head-to-head trials directly compared AI systems, indirect comparisons revealed no differences between AI platforms. Conclusions: Current evidence supports size-dependent efficacy of AI-assisted colonoscopy, with maximal benefit for diminutive polyps (≤5mm) and diminishing returns as polyp size increases. EndoScreener demonstrates the most consistent evidence for improving diminutive lesion detection. These findings support targeted AI implementation for screening populations where small lesion detection impacts surveillance intervals. Direct comparative trials between AI systems are needed to establish optimal platform selection. Clinical Trial: The protocol was registered with PROSPERO, registration number: CRD420251266932.

  • Abstract NHS England's early digital-first agenda has been absorbed into a broader reform of modern general practice built around multiple access routes, structured triage, and routine use of online consultation systems. The central question is therefore no longer whether digital access should exist, but under what conditions it improves care. This review reassesses digital-first primary care in England through the linked lenses of access, equity, communication, continuity, and trust. Recent UK evidence shows that digital routes can reduce friction for routine and bounded problems, improve administrative convenience, and widen options for contacting general practice. However, these gains are conditional rather than universal. Patients with limited digital confidence, language barriers, multimorbidity, or needs that are diagnostically or relationally complex are less well served when digital systems become the default route rather than one option within a broader access model. New English studies also show that patients judge access less by speed alone than by whether they can reach an appropriate professional in an acceptable mode and, where possible, maintain continuity with a known clinician. At the same time, staff studies indicate that digital access reform can redistribute work, create hidden facilitation labour, and strain relationships within practice teams. We argue that the most useful analytic and policy endpoint is no longer digital-first primary care, but patient-centred hybrid care. In this review, that term denotes a multi-channel model in which route of access, consultation modality, continuity, and safety are deliberately matched to patient need rather than subordinated to technological throughput. Future research should prioritize subgroup-sensitive outcomes, continuity, trust, safety, and implementation burden.

  • Background: National surveillance of commercial retail environments is limited by data sources that are updated infrequently and capture narrow dimensions of food access. Google Places API offers continuously updated, programmatically accessible data on business locations across the US, but its validity as a population-level exposure measure has not been systematically evaluated. Objective: To develop and evaluate a Google Places–derived US Built Retail Environment Index (UBER) as a scalable measure of county-level commercial retail infrastructure in the contiguous United States and to estimate its spatial association with age-adjusted diabetes prevalence. Methods: We constructed the US Built Environment Retail (UBER) Index using Google Places API data capturing alcohol outlets, fast-food and convenience stores, grocery and health food stores, and fitness and recreation facilities across 1,701 US counties. Principal component analysis generated a single composite index. We assessed Construct validity against USDA Food Access Research Atlas and County Health Rankings benchmarks and estimated spatial associations with age-adjusted diabetes prevalence (CDC PLACES, 2025) using spatial error models adjusting for area deprivation, urbanicity, and census division. Quantile regression and tract-versus-county comparisons evaluated robustness. Results: The first principal component explained 92.9% of variance among the four indicators with near-equal loadings (range: 0.492–0.506), indicating the index captures overall commercial retail density rather than any single establishment type. Correlations with existing food-environment benchmarks were weak (|r| < 0.20), confirming that the index measures a distinct dimension of the built environment not reflected in current surveillance tools. Each SD increase in the UBER Index was associated with 0.24 percentage points higher diabetes prevalence (95% CI, 0.16–0.32; P < .001). Associations were stable across quantiles of diabetes prevalence. County- and tract-level analyses showed discordant effect directions, indicating scale-dependent ecological confounding. Conclusions: Google Places API data can generate a reliable, spatially structured index of commercial retail infrastructure associated with county-level diabetes prevalence. The UBER index captures a dimension of the built environment not represented in existing surveillance systems and may support scalable digital monitoring of commercial environments relevant to chronic disease prevention.

  • Digital Delivery of Lifestyle Interventions in Online Clinical Trials: An Umbrella Scoping Review

    Date Submitted: Mar 23, 2026
    Open Peer Review Period: Mar 24, 2026 - May 19, 2026

    Background: Noncommunicable diseases (NCDs) cause a very significant health and economic burden. As they are associated with modifiable behavioral risk factors such as physical inactivity and poor diet, more evidence on effective health behavior modification methods is needed. Fully online delivery of clinical trials can provide a practical and scalable way to evaluate interventions that aim to modify relevant lifestyle factors. The emergence of online delivery methods presents opportunities and challenges that need to be better understood to inform future research. Objective: This umbrella scoping review aimed to summarize current evidence on the opportunities and challenges provided by lifestyle intervention clinical trials that are delivered fully online. Methods: Evidence was synthesized from existing peer-reviewed review papers to map the digital delivery methods in online lifestyle intervention trials, focusing on technologies, recruitment, engagement and retention strategies, and reported strengths, limitations, and future directions. Using PRISMA-ScR guidelines, PubMed, EMBASE, CINAHL, Web of Science, and Scopus were searched for reviews published between January 2013 and May 2025. Predominantly (>50%) hybrid, telehealth, or acute condition focused interventions were excluded. Results: Eligible reviews (n=39) discussed digital interventions targeting diet, physical activity, or both, for lifestyle improvement, chronic disease prevention or management. The most common hardware used in online lifestyle clinical trials were smartphones and wearables, with the most frequent software modes being web-based platforms, mobile apps, and SMS. Successful engagement strategies often integrated behavior changes techniques, such as goal setting, self-monitoring, personalized feedback, and human support into the intervention design, or had behavior change techniques as a feature of the technology itself. Reported strengths of conducting clinical trials online included improved accessibility, scalability, cost-efficiency and personalization, whereas limitations discussed were poor engagement and retention, digital literacy barriers, and rapid technological change outpacing evaluation capabilities. Interventions that used theory-based designs, particularly those using Social Cognitive Theory, the Transtheoretical Model, and the Theory of Planned Behavior, were reportedly most successful in improving behavioral outcomes. Engagement and retention varied considerably across online trials, suggesting that the success of these studies may depend less on the online delivery modality itself and more on how interventions and technologies are designed, including the integration of behavioral theory and behavior change techniques. Conclusions: This review shows that online delivery of lifestyle intervention trials is a feasible and potentially advantageous method as it can improve reach, increase scalability, be cost-efficient, and allow more personalization of the intervention. To further improve the conduct of online clinical trials, future research should address increased use of behavioral change theory, equitable access to clinical trial participation, management of data privacy and security, intervention fidelity, and use of novel technologies such as artificial intelligence in a field that is rapidly evolving. Clinical Trial: The protocol was prospectively registered with the Open Science Framework https://osf.io/umcfv; 5 September 2023

  • Background: Clinical guidelines recommend an integrated, person-centered care model with better control of modifiable risk factors and coexisting conditions in patients with atrial fibrillation (AF), but many persons with AF receive insufficient risk factor management. Digital health technologies may provide valuable support in addressing this gap. Objective: Our aim was to evaluate a co-designed digital platform for supporting person-centered management of modifiable risk factors in individuals with AF. Methods: This is a mixed-methods study including a standardized quantitative questionnaire used to score the usability of digital tools, the System Usability Scale (SUS), and a qualitative, descriptive, manifest content analysis of individual interviews. Results: Twenty-two patients hospitalized for AF were included (age 68 (48-79) years; 32% female; BMI 27.7 (20.8-35.0) kg/m2; paroxysmal/persistent AF (36%/64%); AF duration 4 (0.5-18 years). Relevant comorbidities were hypertension (77%), heart failure (36%), diabetes mellitus type 2 (14%), and ischemic heart disease (18%). Usability was rated high, with a mean SUS score of 75 (±18.2), indicating above-average user acceptance. Participants’ requirements were summarized into four main categories and ten subcategories. First, they value a clear layout with simple design and easy navigation. Second, they appreciate positive content, which is informative, inclusive and motivating. Third, they request personalized information on different aspects and provided on different levels. Fourth, they desire individualized medical recommendations that are personalized and flexible but open to individual choice. Conclusions: To improve digital management of lifestyle-related risk factors and comorbidities, individuals with AF seek a solution with a clear layout, positive content, personalized information, and individualized medical recommendations.

  • Data Missingness in Digital Phenotyping and Its Implications for Clinical Inference and Decision Making: the TRIM framework

    Date Submitted: Mar 19, 2026
    Open Peer Review Period: Mar 22, 2026 - May 17, 2026

    Background: Digital phenotyping — the use of personal digital devices to capture real-world behavioral and physiological data — holds promise for measuring patient function over time. However, missing data remains a pervasive challenge in longitudinal studies where missingness may obscure clinically relevant signals. Existing statistical methods focus on group-level inference and offer limited guidance on how causes of missingness should inform clinical interpretation and decision-making. Objective: This study aimed to (1) characterize data missingness across multiple timescales in a longitudinal digital phenotyping study of patients with chronic pain, (2) evaluate the effect of imputation on clinical inference, and (3) propose a practical framework for categorizing and responding to missing data in clinical digital phenotyping research. Methods: We analyzed data from 85 patients with chronic musculoskeletal pain (mean age 55.2 years, SD 15.7; 51 female, 32 male, 1 transgender) recruited from the Pain Intervention and Digital Research Program. Active data (PROMIS-29 surveys, daily pain scores) and passive data (accelerometer, GPS) were collected via the Beiwe Research Platform over 180 days. Data completeness was computed at day, hour, and minute levels. Linear mixed-effects models assessed associations between daily missingness and Forest-derived summary measures (cadence, home time, significant locations). Cumulative Link Mixed Models and linear mixed models with autoregressive error structures evaluated associations between PROMIS domain scores and digital measures, adjusting for age, sex, race, and within-subject correlation. Complete-case analyses were compared against multiple imputation (predictive mean matching, proportional odds logistic regression, and MidasTouch) using 100 imputed datasets combined via Rubin's rules. Results: Median accelerometer completeness was 60% at the day level, 37% at the hour level, and 26% at the minute level; GPS completeness followed a similar pattern (57%, 34%, and 5%, respectively). Cadence showed no significant association with missingness (false discovery rate [FDR]-adjusted P=.32). The number of significant locations declined modestly with higher missingness (beta=-0.073 per 10 percentage points; FDR-adjusted P<.001). In complete-case analysis, higher cadence was associated with lower depression scores (95% CI -2.09 to -0.28; P=.01); this association was attenuated after multiple imputation (P=.13). Older participants remained enrolled longer (hazard ratio 0.979 per year; P=.02) but were less engaged while enrolled (odds ratio 0.962; P=.004). Race and sex were not significantly associated with engagement or retention. Conclusions: Data missingness in digital phenotyping varies substantially by the timescale at which it is assessed, and imputation choices can alter clinical interpretations. We propose the Triage and Response for Interpreting Missingness (TRIM) framework, which categorizes causes of missing data into Technology Failure, Clinically Relevant events, and Extraneous life events, each requiring distinct operational and analytical responses depending on the clinical context. TRIM provides a shared vocabulary for clinicians, statisticians, and engineers to ensure that missing data is meaningfully interpreted rather than merely imputed.

  • AI for Sexually Transmitted Infection Detection: A Call for Robustness, Ethical Oversight, and Equitable Deployment

    Date Submitted: Mar 16, 2026
    Open Peer Review Period: Mar 18, 2026 - May 13, 2026

    The application of artificial intelligence (AI) in medicine has dramatically changed the paradigm of clinical practice. Several studies recently published in ‘Journal of Medical Internet Research’ have witnessed this comprehensive evolution which spans from diagnosis, treatment to prevention. However, our experience as clinicians and digital health researchers suggests critical, underexplored facets in the process of using AI technology to advance the diagnosis and treatment of sexually transmitted infection (STI) diseases, including the inherent heterogeneity in real-world image acquisition, the societal and ethical ripple effects of STI, and the successful and equitable integration of advanced computational tools in diagnostic pathways. On this basis, we also envisioned a framework that clinicians should follow within the AI-assisted diagnostic process. Collectively, the above issues are worth warranting further attention to enhance the true clinical utility, interpretability, and equitable deployment of such technologies.

  • A Multilingual Digital Micro-Learning Intervention for Oral Health in Refugee Shelters: Randomized Controlled Trial

    Date Submitted: Mar 17, 2026
    Open Peer Review Period: Mar 17, 2026 - May 12, 2026

    Background: Refugees frequently face language and access barriers to preventive health information, contributing to avoidable oral disease and widening health inequities. Scalable digital micro-learning interventions may offer a pragmatic strategy to reduce health information barriers in settings where conventional health education is difficult to deliver. y Objective: This randomized controlled trial aimed to evaluate whether a brief multilingual digital micro-learning intervention improves clinical oral hygiene outcomes and selected oral health literacy–related outcomes compared with delayed access control among adults living in refugee shelters. Methods: A two-arm, parallel-group randomized controlled trial was conducted among 86 adults living in two municipal refugee shelters in Germany (November 2023 - February 2024). Participants were randomized (1:1) to a four-minute animated oral-hygiene micro-learning video (“GlobeSmile”; Arabic, Kurdish, Dari, Farsi, Ukrainian) or delayed access. The intervention was viewed once at baseline on a tablet, with optional re-viewing via QR code for two months. Primary outcomes were change in Plaque Index (PI) and Gingival Index (GI) from baseline to two-month follow-up. Secondary outcomes included oral health knowledge, attitudes, perceived self-efficacy, and self-reported behaviours. Results: Eighty-three of 86 participants completed follow-up (97%). PI decreased more in the intervention group than in controls (mean change −0.21 ± 0.27 vs −0.04 ± 0.17; p = 0.002). GI decreased in both groups without significant between-group difference. Questionnaire findings indicated improvements in selected knowledge, attitude, and self-efficacy outcomes in the intervention group, while most self-reported behaviours and lifestyle factors changed little. Approximately three-quarters of intervention participants reported re-viewing the video at least once, and re-viewing was associated with larger PI reductions. Conclusions: Brief multilingual digital micro-learning may improve plaque control and selected oral health literacy outcomes in refugee shelters, supporting its potential as a scalable, low-cost strategy to reduce health information barriers in underserved settings. Clinical Trial: German Clinical Trials Register (DRKS00032017).

  • Training AI Models for Aesthetic Facial Evaluation: A Focused Review and Framework to Mitigate Homogenizing Bias

    Date Submitted: Mar 16, 2026
    Open Peer Review Period: Mar 17, 2026 - May 12, 2026

    As artificial intelligence (AI) models become increasingly integrated into facial aesthetic surgery for attractiveness prediction and surgical outcome simulation, their potential to perpetuate bias poses clinical concerns. Current models trained on limited datasets inaccurately evaluate underrepresented populations and risk promoting aesthetic homogenization that conflicts with patient goals of ethnic feature preservation. Drawing on current literature, this paper examines bias across AI development stages in aesthetic facial evaluation. Benchmark datasets such as SCUT-FBP and Chicago Face Database underrepresent elderly, non-White, and ethnically diverse populations. Training methodologies lack fairness-aware techniques, and evaluation focuses on overall rather than demographic-stratified accuracy. While individual mitigation strategies exist—including balanced datasets, adversarial debiasing, and fairness metrics—no comprehensive framework integrates these approaches across the entire development lifecycle. We propose a six-pillar framework spanning the AI development lifecycle: (1) diverse data collection with synthetic augmentation, (2) fairness-aware training techniques, (3) complementary fairness metrics with intersectional assessment, (4) explainable AI for clinical transparency, (5) stakeholder engagement, and (6) continuous monitoring. Despite the challenges of maintaining algorithmic standardization and cultural specificity, this framework provides implementation guidance for AI developers, clinicians, and institutions, with principles applicable beyond aesthetic surgery to broader facial analysis applications.

  • Governing the Digital Health Commons: Ostrom's Design Principles for Strategic Planning

    Date Submitted: Mar 16, 2026
    Open Peer Review Period: Mar 17, 2026 - May 12, 2026

    Digital health transformation has absorbed tens of billions of dollars in public investment over two decades and produced a consistent pattern of failure. The United States spent $36 billion on electronic health record incentives that achieved adoption without interoperability. England's National Programme for IT was cancelled after a decade and £13 billion. Australia's My Health Record required twelve years and legislative compulsion before providers shared data. These failures have been attributed to management shortcomings, funding gaps, and political interference, yet none of these explanations accounts for why the same pattern recurs across countries with different political systems, funding models, and vendors. This paper proposes that the cross-national consistency reflects a structural governance problem. Drawing on Elinor Ostrom's polycentric governance framework, the analysis argues that digital health ecosystems function as commons governed by multiple autonomous decision centers with overlapping jurisdictions. When Ostrom's eight design principles are absent, coordination fails predictably regardless of technology or funding. The paper tests this claim against six initiatives across three countries, mapping each against the design principles and identifying which are systematically absent. A contrast case, OpenNotes, demonstrates that partial presence of these principles improves outcomes. The analysis translates these findings into a governance-first strategic planning framework specifying requirements for boundary definition, stakeholder inclusion, outcome monitoring, graduated sanctions, conflict resolution, and nested governance. For digital health strategic planning, the contribution is foundational: hierarchical planning frameworks are structurally inadequate for polycentric environments, and Ostrom's design principles offer a rigorous alternative.

  • Safety of telemedicine versus in-person care for patients with tracheal devices: A propensity score–matched cohort study

    Date Submitted: Mar 16, 2026
    Open Peer Review Period: Mar 17, 2026 - May 12, 2026

    Background: Patients with tracheal diseases often require long-term follow-up after tracheal device placement, with a risk of adverse events that may lead to emergency care and unplanned interventions. Telemedicine has been proposed as an alternative to in-person follow-up to improve access and continuity of care. Objective: The primary objective was to compare the need for emergency department visits between telemedicine and in-person groups. Secondary objectives included comparing hospital readmissions, 30-day hospital readmissions, and unplanned interventions between groups. Methods: This retrospective, single-institution study included adult patients with tracheal devices who underwent telemedicine and in-person outpatient clinic visits between 2020 and 2024. To balance the groups, we used 1:1 propensity score matching. We collected demographic and clinical data and evaluated the need for emergency department visits, hospital readmissions, 30-day hospital readmissions, and unplanned interventions. Kaplan–Meier estimation of time to first emergency department visit was performed to assess outcomes after outpatient visits. Results: A total of 483 patients (277 telemedicine and 206 in-person) underwent 2487 visits (1258 telemedicine and 1229 in-person). After propensity score matching, 336 patients remained (168 in each group). There were no significant differences in the need for emergency department visits, hospital readmissions, or unplanned interventions. Telemedicine group had significantly fewer 30-day hospital readmissions (OR = 0.38; 95% CI = 0.16–0.87; p = 0.021). Kaplan–Meier analysis indicated no statistically significant difference in emergency department–free visits. Conclusions: A telemedicine outpatient program is safe for the follow-up of adult patients with tracheal devices.

  • Care Robots as Emerging Health Technologies: A Systematic Review and Meta-Analysis

    Date Submitted: Mar 13, 2026
    Open Peer Review Period: Mar 16, 2026 - May 11, 2026

    Background: The rapid growth of aging populations worldwide is placing mounting pressure on long-term care systems already facing critical nursing workforce shortages. Care robots—including socially assistive robots (eg, PARO, NAO, Pepper), companion robots, and therapeutic robots—have emerged as a promising category of digital health technology designed to complement professional caregiving in older adult populations. Despite increasing deployment across clinical and community care settings, a comprehensive quantitative synthesis of their clinical effectiveness across diverse patient populations and outcome domains has not been established. Objective: This systematic review and meta-analysis aimed to quantify the pooled effects of care robot interventions across key patient outcome domains, identify moderating factors including robot type and target population characteristics, appraise study methodological quality, and evaluate the overall certainty of evidence to inform clinical implementation decisions. Methods: We searched PubMed, CINAHL, Cochrane Library, and Embase from inception through December 2024 following PRISMA 2020 guidelines. Eligible studies were randomized controlled trials (RCTs) or quasi-experimental studies comparing care robot interventions with standard care, active controls, or waitlist controls in any clinical or community setting. Methodological quality was assessed using the Cochrane Risk of Bias 2 (RoB 2) tool; certainty of evidence was evaluated using the GRADE framework. Pooled standardized mean differences (SMDs; Hedges' g) with 95% CIs were calculated using random-effects models. Results: Thirty-four studies (N = 2,476 participants; 17 countries; 2015–2024) met inclusion criteria; 28 were included in meta-analyses. Care robot interventions yielded statistically significant effects compared with controls for neuropsychiatric symptoms (k = 8; SMD = −0.34; 95% CI −0.62 to −0.06; I² = 64%), quality of life (k = 6; SMD = 0.27; 95% CI 0.03 to 0.51; I² = 52%), agitation (k = 5; SMD = −0.31; 95% CI −0.55 to −0.07; I² = 48%), stress and pain (k = 6; SMD = −0.38; 95% CI −0.68 to −0.08; I² = 72%), and social-communicative skills (k = 6; SMD = 0.45; 95% CI 0.14 to 0.76; I² = 54%). Effects on depression and cognitive function were not statistically significant. Subgroup analyses indicated that PARO demonstrated the strongest effects in older adults with dementia, whereas humanoid robots (NAO) yielded the largest effect sizes for children with autism spectrum disorder (ASD). GRADE certainty of evidence ranged from moderate (neuropsychiatric symptoms, agitation) to very low (depression, cognitive function). Conclusions: Care robots represent a viable and evidence-supported category of digital health technology with significant effects across multiple patient outcome domains relevant to aging care. These findings support integration of care robots—particularly PARO in dementia and long-term care settings—as complementary digital health interventions. Successful implementation requires attention to technology acceptance, facilitator training, and interoperability with existing health information systems. Clinical Trial: Not applicable

  • Co-developing an implementation plan for a digital health intervention: an example in rural Missouri clinics

    Date Submitted: Mar 13, 2026
    Open Peer Review Period: Mar 16, 2026 - May 11, 2026

    Background: Given the rapid growth of digital healthcare, greater transparency in engaging end-users in developing implementation plans is needed to improve the sustained delivery and effectiveness of digital healthcare. PREVENT is a patient-centered digital health tool designed to engage patients with overweight or obesity in conversations about behavior change to lower their cardiovascular disease (CVD) risk. We are implementing PREVENT in a rural federally qualified health center network with eight clinics. Objective: Objective: The objective of this paper is to present a multi-method approach used to engage clinic teams, including healthcare team members and administrators, to co-develop an implementation plan for the PREVENT digital health tool. Methods: Methods: Healthcare team members (e.g., clinicians, community health workers) and administrators (e.g., informatics professionals, clinic managers, CEO) were engaged through qualitative interviews, advisory board meetings, and site visits to develop an implementation plan prior to implementation. Qualitative interviews were coded using the Consolidated Framework for Implementation Research to identify potential barriers and facilitators of implementing PREVENT and paired with learnings from previous trials of PREVENT to inform the selection and tailoring of implementation strategies. Direct clinic observations generated clinic workflow maps, roles and responsibilities of team members, clinic resources (e.g., technology, space), and an understanding of the clinic and community context and readiness for implementation. A mixed-methods evaluation plan was developed using validated measures, input from the advisory board, and implementation research logic models. Results: Results: Implementation of the PREVENT tool was facilitated by a positive clinic culture, alignment with chronic disease priorities, an easy-to-use and adaptable design with EHR integration, and CHWs’ experience addressing social needs. Key barriers included limited technology infrastructure, variable staffing across clinics, patients’ digital access and literacy challenges, and the need for additional training and support for care team members. An implementation plan that includes thirteen strategies across six implementation strategy clusters was created to leverage facilitators and address barriers. A multi-modal evaluation plan was created to examine implementation and effectiveness. Conclusions: Conclusions: This paper provides an example of how to develop implementation and evaluation plans tailored to the healthcare context that engage end users to increase the impact of a digital health intervention. This work may be replicated to support the successful implementation of other digital health tools, particularly in under-resourced, complex healthcare contexts such as rural clinics where resources are less available.

  • Background: The prevalence of mental health disorders is steadily increasing, while informative biomarkers remain lacking. Although artificial (AI) intelligence shows promise for revealing latent patterns in data, available datasets in the computational psychiatry community are still insufficient. Digital technologies and informatics tools could fill this gap, offering new strategies for collecting large real-world data. Furthermore, computational infrastructures provide scientists with access to e-services and powerful computational resources. Objective: We present the NewPsy4U web platform, which integrates data, AI pipelines, and mobile applications into an efficient environment from the end-user’s perspective. Methods: NewPsy4u is built on a LAMP architecture (Linux, Apache, MySQL, PHP) and is hosted at the IRCCS-FBF High Performance Computing (HPC) Center. These resources enable the execution of classical and generative AI algorithms required for computational psychiatry. The platform implements the highest standard protocols, while access is granted following registration and approval. The data repository complies with the FAIR principles (Findable, Accessible, Interoperable, Reusable), and the datasets provided are structured according to the OMOP Common Data Model or BIDS standard, enabling standardized data storage and interoperability. Mobile technology is integrated into NewPsy4u, allowing users to collect patient data using the experience sampling method (ESM). All mobile data are synchronized within the NewPsy4u web-based portal, where they can be managed for research purposes. Results: NewPsy4u is designed to provide access to pipelines and multiple datasets. The platform is freely accessible and its AI algorithms are released as open-source tools to promote transparency, reproducibility, and collaborative development. All AI algorithms available on the platform are offered as group or single-case second opinion tools. Multimodal patient records can be hosted in the web-platform with either open or restricted access. NewPsy4u currently includes data from 1900 patients diagnosed with various psychiatric conditions, collected at multiple time points. The platform architecture supports integration of multimodal data (sociodemographic, clinical, imaging and digital information captured by a mobile app) and serves as both a data-sharing solution and a hypothesis testing environment. Conclusions: NewPsy4u is a platform developed to support both research and clinical settings, offering an integrated suite of digital tools for psychiatry.

  • The Digital Twin Paradigm in Head and Neck Cancer: Clinical Opportunities and Challenges

    Date Submitted: Mar 13, 2026
    Open Peer Review Period: Mar 16, 2026 - May 11, 2026

    Head and neck squamous cell carcinoma (HNSCC) represents a significant public health challenge. its treatment is characterized by high anatomical complexity and a critical need for functional preservation of speech, swallowing, and respiration. While multimodality care, including surgery, radiotherapy, and systemic therapy, has substantially improved patients’ survival, there are still high rates of life-altering toxicities and low response rates to immunotherapy. Current clinical decision-making largely relies on "digital snapshots"—static representations derived from population-based statistics and time-point-specific imaging—which usually fail to account for the rapid anatomical and biological change during treatment course. Digital twins hold great promise in accelerating scientific breakthroughs and transforming oncology treatment and precision medicine by the creations of high‑fidelity virtual patient representations that integrate real‑time biological, anatomical, and clinical data. In this viewpoint, we propose three interconnected digital twins to establish a conceptual framework for the HNSCC digital twin. Those digital twins include the anatomical twin, utilizing virtual surgical planning and augmented reality for geometric precision; the dosimetric twin, employing daily imaging and synthetic CT generation for automated adaptive radiotherapy; and the biological twin, integrating deep phenotyping, radiomics, and mechanistic omics to simulate individualized therapeutic trajectories. Beyond acute care, we explore the clinical utility of digital twins in navigating the "de-escalation dilemma" for HPV-associated disease and improving survive through digital phenotyping. With the advent of longitudinal data from medical-purpose Internet of Things (IoT) sensors and circulating tumor DNA (ctDNA) kinetics, digital twins can serve as active sentinels, detecting functional decline or subclinical events months before clinical manifestation. Furthermore, the implementation of synthetic control arms via virtual populations offers a venue to accelerate clinical trials and address ethical issues in rare molecular subtypes. Despite this potential, substantial barriers to implementation remain, including profound anatomical instability during treatment, the "black box" nature of deep learning algorithms, and the challenges of multiscale data integration. We provide a technical roadmap for the development of "morphing" digital twins—dynamic systems capable of continuous geometric and biological auditing. By bridging the gap between the macroscopic world of radiation physics and the microscopic world of cellular evolution, digital twins promise to shift the management of HNSCC from statistical probability toward personalized biological determinism, optimizing HNSCC management while maintaining the essential human functions of the patients.

  • Patterns, Perceptions, and Professional Use of Social Media Among Healthcare Professionals: A Mixed-Methods Study

    Date Submitted: Mar 15, 2026
    Open Peer Review Period: Mar 15, 2026 - May 10, 2026

    Background: Social media is increasingly used by healthcare professionals for professional communication, information sharing, and education. However, how clinicians integrate social media into their scientific information practices, and how they perceive its risks and benefits, remains incompletely understood across diverse settings. Objective: This study aimed to examine patterns of social media use among healthcare professionals, identify factors associated with scientific information seeking and medical content posting, and explore perceptions of social media through integrated quantitative and qualitative analyses. Methods: During the month of May 2025, we conducted an international, cross-sectional, mixed-methods survey of healthcare professionals, including physicians and paramedical staff. The questionnaire was administered electronically using Google Forms and disseminated via newsletters and targeted email invitations to healthcare professionals through established professional and academic networks. Multivariable logistic regressions were performed to identify factors associated with social media use for scientific information and having ever posted medical information on social media. Adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were reported. Model estimates were evaluated for plausibility and precision. A multinomial logistic regression was performed to identify factors associated with the risk perception of social media. Outcomes included use of social media for scientific information, posting of medical content, and perceptions of social media as an opportunity, risk, or both. Free-text responses were analysed using inductive thematic analysis, and findings were triangulated with quantitative results. Results were reported as relative risk ratios (RRRs) with 95% confidence intervals. All quantitative analyses were conducted using Stata (15, StataCorp LLC, College Station, TX), and statistical significance was assessed at a two-sided α level of 0.05. Results: A total of 1,204 respondents participated. Social media use was frequent, with 67.5% reporting daily use and 87.2% reporting professional use. Bibliographic databases (47.6%) and scientific journals (22.1%) remained the primary sources of scientific information, while 8.6% identified social media as their primary source. Use of social media for scientific information was strongly associated with professional social media use (adjusted OR 7.40, 95% CI 4.87–11.25). Posting medical information was more common among male respondents and frequent social media users, and less common among older professionals and nursing or allied health staff. Most respondents perceived social media as both an opportunity and a risk (60.8%). Qualitative findings highlighted accessibility, time efficiency, and habit as drivers of engagement, alongside concerns about misinformation, professionalism, and content quality. Conclusions: Social media is primarily viewed as a complementary tool whose value depends on critical use and professional norms, whereas healthcare professionals continue to rely on traditional scientific sources for core knowledge acquisition. Strengthening guidance, digital literacy, and institutional support may help maximise benefits while mitigating risks.

  • Background: Attention-deficit/hyperactivity disorder (ADHD) is a prevalent neurodevelopmental condition in children and adolescents, for which conventional treatments present certain limitations. While digital therapeutics (DTx) have developed rapidly, the relative efficacy of different DTx modalities for this population remains to be established. Objective: To systematically compare the efficacy of four digital therapeutics (DTx) modalities (single-task, cognitive-motor dual-task, AI-integrated single-task, and AI-integrated cognitive-motor dual-task) on core symptoms and executive functions in children and adolescents with ADHD within the dual framework of task design and AI empowerment. Methods: We systematically searched PubMed/MEDLINE, PsycINFO, Web of Science, EMBASE, Scopus, ProQuest Dissertations and Theses, the Cochrane Library, and grey literature from ClinicalTrials.gov for randomized controlled trials published from January 2000 to February 2026, without language restrictions. A snowballing method was also employed. Risk of bias was assessed using the Cochrane RoB 2.0 tool. Data were analyzed using Bayesian network meta-analysis in R software (version 4.2.3). Heterogeneity was assessed using I² statistics, and publication bias was evaluated using Egger's test. Subgroup analyses, meta-regression, and sensitivity analyses were performed to explore sources of heterogeneity. Results: A total of 32 studies involving 2,819 patients were included. Network meta-analysis showed that AI-integrated cognitive-motor dual-task DTx appeared to be the most effective modality for improving core symptoms and executive functions, demonstrating the highest probability of being the best treatment on the Attention Deficit/Hyperactivity Disorder-Rating Scale (ADHD-RS) [Surface Under the Cumulative Ranking Curve (SUCRA): 57.5%; Mean Difference (MD): -3.03, 95% Confidence Interval (95% CI): -5.59 to -0.47], the Swanson, Nolan, and Pelham Rating Scale, Version IV - Inattention subscale (SNAP-IV-PI) [SUCRA: 58%; MD: -5.58, 95% CI: -8.76 to -2.39], the Swanson, Nolan, and Pelham Rating Scale, Version IV - Hyperactivity-Impulsivity subscale (SNAP-IV-PHI) [SUCRA: 81.6%; MD: -6.84, 95% CI: -10.37 to -3.31], and the Behavior Rating Inventory of Executive Function (BRIEF) [SUCRA: 67.4%; MD: -7.75, 95% CI: -10.06 to -5.43]. Moreover, this modality significantly outperformed conventional pharmacotherapy across all outcomes. Subgroup analyses revealed that intervention duration emerged as a potential source of heterogeneity for the SNAP-IV (both PI and PHI subscales) and BRIEF, while mean participant age was identified as a potential source of heterogeneity for the SNAP-IV-PI and BRIEF (all P < 0.05). Sensitivity analyses indicated that individual studies influenced heterogeneity. Of note, all outcome measures reported were based on parent versions of the scales. Conclusions: AI-integrated cognitive-motor dual-task DTx may be the most effective intervention for improving core symptoms and executive functions in children and adolescents with ADHD. Subgroup analyses suggested that Intervention duration and age emerged as moderators of treatment outcomes, warranting consideration in clinical practice. Clinical Trial: CRD420261304236

  • Background: Health communicators and researchers are increasingly exploring partnerships with social media influencers as a strategy to improve the reach and engagement of health messaging. However, practical guidance on how health communicators can identify, recruit, and collaborate with influencers is limited. Objective: The aim of this paper is to provide a detailed description of how to work with social media influencers to disseminate health messages and to highlight lessons learned that may help others overcome challenges associated with this communication channel. Methods: We conducted a process evaluation of an Instagram influencer campaign promoting colorectal cancer screening between March and July 2025. We reviewed publicly available guidance on collaborating with social media influencers for health promotion and summarized key recommendations. Using a paid influencer marketing platform, we identified and contacted nano- and micro-influencers (1,000–50,000 followers). Participating influencers created and posted an Instagram Reel and shared it to their Instagram Story. We documented the recruitment process, vetting criteria, negotiations, content review procedures, and engagement metrics and examined associations between influencer characteristics and engagement outcomes. Results: We sent 1,907 outreach emails to potential influencers; 72 expressed interest and we negotiated terms with 52 before finalizing agreements with 22 and receiving Reels from 16 (0.84%). Outreach emails that specified compensation and project details upfront were the most effective strategy (completed posts from 2.0% of outreach emails compared with 0.7% for emails that did not include compensation and 0% for emails sent only after pre-vetting influencers). Recruiting these influencers required approximately 2.5 months of outreach and a paid influencer marketing platform subscription costing $1,647. Influencer payments ranged from $200 to $500 (mean $389). The 16 influencer videos generated 89,764 total views (mean 5,610 per video) and approximately 232 visits to the campaign website. We found no significant associations between influencer payment or follower count and video views or engagement rates. Conclusions: Partnering with influencers to disseminate health messages on social media can result in relatively high engagement with health messages, including among audiences who may not actively seek health information. However, implementing influencer campaigns using a commercial influencer marketing platform required substantial recruitment effort, including large volumes of outreach and lengthy negotiation timelines. In our campaign, fewer than 1% of outreach emails resulted in completed posts. Providing compensation and project expectations in the initial outreach email substantially improved recruitment success. Influencers with relatively fewer followers may generate similar reach and engagement at lower cost, but may be less experienced and require more clarifications in the negotiation process. Establishing clear expectations for deliverables and revisions may help prevent delays and improve content quality. The guidance in this study can help health communicators develop more realistic implementation plans and budgets when considering influencer-based health communication campaigns.

  • Background: Chronic primary pain is a complex condition involving biological, psychological, and behavioral mechanisms and is commonly associated with emotional distress and reduced quality of life (QoL). Digital mental health interventions (DMHIs) offer scalable and accessible solutions for delivering psychological care in chronic pain management; however, evidence regarding their effectiveness across delivery modalities and outcome domains remains heterogeneous. Objective: This systematic review aimed to (1) evaluate the effectiveness of DMHIs on clinical (pain intensity, disability) and psychological outcomes (QoL, anxiety, depression, catastrophizing, and self-efficacy) in adults with chronic primary pain; (2) examine whether specific digital delivery modalities are differentially associated with particular outcomes; and (3) identify methodological gaps to inform future research and implementation. Methods: A systematic literature search was conducted in PubMed, Scopus, PsycINFO, Cochrane Library, Web of Science, and Google Scholar following PRISMA guidelines. Two independent reviewers screened randomized controlled trials (RCTs) and assessed risk of bias using the Cochrane Risk of Bias 2.0 tool. Given substantial heterogeneity in study designs, interventions, and outcome measures, a narrative synthesis was performed. Results: Twenty-two RCTs were included. DMHIs were effective in improving psychological functioning and pain-related disability, often independently of changes in pain intensity, particularly when grounded in evidence-based psychotherapeutic frameworks such as cognitive behavioral therapy and acceptance and commitment therapy. Guided web-based interventions demonstrated the most consistent benefits, whereas unguided interventions showed smaller effects. Mobile applications and virtual reality–based interventions also showed positive effects on emotional functioning, self-management, and pain interference. Interventions incorporating some form of human guidance were generally associated with superior outcomes. Conclusions: DMHIs represent a promising, scalable, and person-centered approach to improving psychological well-being and functional outcomes in adults with chronic primary pain, particularly when integrated into stepped-care or hybrid care models. Clinical Trial: CRD420251010767

  • Background: Online patient reviews are widely used by consumers to assess the quality of direct-to-consumer teleconsultation (DTCT) services, particularly in settings where objective quality information is limited. However, whether these reviews validly reflect actual clinical and patient-centered care quality remains unclear. Objective: This study aimed to evaluate the validity of online physician reviews in reflecting the quality of care delivered on China’s three largest DTCT platforms. Methods: We conducted a cross-sectional study using unannounced standardized patients (USPs) to objectively assess the quality of DTCT services. Thirty-three USPs were trained to present 11 standardized clinical cases and completed 542 DTCT consultations between physicians on three major Chinese platforms. Technical quality was assessed using a clinical guideline adherence checklist, and patient-centered quality was measured using the Patient–Patient-Centered Care Chinese version (PPPC-CN) scale. Online review quality was defined as the positive review rate displayed on each physician’s profile. Agreement between online reviews and measured quality was evaluated using Intraclass Correlation Coefficients (ICCs), with additional rank correlation analyses. Results: Of the 542 consultations initiated, 530 were completed and 404 physicians had publicly available review data. Among all encounters, 53.14% (288/542) were phone-based and 46.86% (254/542) were text-based consultations. The median positive review rate was 99.9% (interquartile range [IQR], 99.4%–100%). Median guideline adherence was low (0.16; IQR, 0.08–0.26), and median patient-centered quality was modest (PPPC-CN score 2.1; IQR, 1.98–2.79). Diagnoses were completely correct in 40.92% (196/530) of consultations. Unnecessary examinations occurred in 1.7% of encounters, and medication prescribing was appropriate in 79.04%. The median consultation time was 13 minutes (IQR, 7–64.69), and the median registration fee was 29.9 yuan (IQR, 26.1–39.9). Agreement between positive review rate and guideline adherence (ICC= 0.002; 95% CI, −0.006 to 0.013) and between positive review rate and patient-centered quality (ICC= 0.014; 95% CI, −0.043 to 0.083) was negligible and far below accepted validity thresholds. Correlations between positive review rate and diagnostic accuracy were weak but statistically significant (Spearman ρ = 0.168; Kendall τ = 0.141; both P < 0.05). Limitations include the use of standardized cases rather than real patients and the focus on publicly visible review metrics. Conclusions: Online reviews on major platforms were overwhelmingly positive but showed almost no alignment with actual provider performance. DTCT providers demonstrated low guideline adherence and modest patient-centered quality. More research on improving the review frameworks is urgently needed to fill the gap between patient feedback and service quality. Clinical Trial: The study has been approved by the Southern Medical University Ethics Committee ([2022] No. 013) and registered with the China Clinical Trial Registry (ChiCTR2200062975).

  • Background: Breast cancer has become the most common cancer worldwide, and newly diagnosed breast cancer patients are in particular need of health information. The results of this study will help clinical providers understand the complete process related to online health information behavior among newly diagnosed breast cancer patients and guide them in developing targeted information behavior intervention. Objective: The purpose of this study was to identify online health information behavior patterns in newly diagnosed breast cancer patients. The purpose of this study was to identify the specific characteristics of patients' online health information behaviors using a grounded theory approach, with the goal of ultimately identifying a behavioral pattern that depicts the trajectories of change among these behaviors. Methods: Thirty-two patients with newly diagnosed breast cancer who underwent breast surgery at The First Affiliated Hospital of Anhui Medical University were interviewed semistructurally per procedural grounded theory from August 2021 to May 2022. The data were processed through three-level coding, continuous comparison, and dimensional analysis until theoretical saturation was achieved. This manuscript adheres to the COnsolidated criteria for REporting Qualitative research (COREQ) guidelines. Results: An online health information behavior pattern was identified for newly diagnosed breast cancer patients. The complete online health information behaviors (core categories) of patients during the period from breast cancer diagnosis to treatment included the following aspects: information acquisition behavior, including active information search, avoidance of information, and encounter information; information evaluation behaviors, including information evaluation and no evaluation; information processing behaviors, including information rejection, information storage, and information utilization; and information processing outcomes, including information termination and continuous searching for information. Conclusions: This study identified a pattern of online health information behavior for newly diagnosed breast cancer patients. The findings will help clinical health care professionals understand the complete process related to online health information behavior among newly diagnosed breast cancer patients, discover entry points for interventions for online health information behavior , and develop targeted information behavioral interventions to help patients improve their e-health literacy, enhance their decision-making ability and information utilization effectiveness, and promote their recovery.

  • From Regulation to Implementation: A Regulatory-Aligned Approach to AI-Powered Digital Platforms in the INTERCEPT Innovative Health Initiative (IHI) Project.

    Date Submitted: Mar 5, 2026
    Open Peer Review Period: Mar 10, 2026 - May 5, 2026

    Regulatory misalignment represents a critical barrier to digital health innovation. However, the Innovative Health Initiative Joint Undertaking (IHI-JU) INTERCEPT (GA n. 101194766), an initiative involving European public and private sectors aimed at intercepting Crohn’s disease prior to the manifestation of symptoms-advocates that the development of AI-powered digital platforms with compliance inherently integrated 'by design' is not only achievable but also imperative as the lack of systematic approaches for integrating regulatory requirements creates inefficiencies and delays in market access. In this perspective, we outline INTERCEPT’s strategic plan for an evolutionary digital infrastructure, coupled with AI-driven analytics, which is meticulously aligned with EU 2016/769 General Data Protection Regulation – GDPR, data privacy regulation, medical device regulation (EU 2017/745 Medical Device Regulation) and recommendations from the Medical Device Coordination Group – MDCG of the European Commission), Health Technology Assessment (HTA) criteria, and the recently enacted EU AI legislation (Regulation EU 2024/1689), which entered into force in 2024 with staggered implementation deadlines extending through 2027. Although presently in the start-up phase with clinical activities about to commence, INTERCEPT functions as a practical blueprint. Methods: Our analysis followed a multi-phase methodology comprising three sequential components. First, we conducted systematic regulatory landscape mapping to identify and categorize the most relevant EU frameworks (GDPR, Data Governance Act, EHDS, Clinical Trials Regulation, AI Act, and MDR) applicable to digital health innovation. Second, we implemented a structured stakeholder engagement through comprehensive surveys administered to all Work Package leaders (n=10), assessing both regulatory relevance and anticipated regulatory risk using high/medium/low categories, coupled with implementation timeline mapping (2025-2029). Third, we established iterative feedback cycles through structured dialogue sessions with project coordinators to identify current and anticipated regulatory challenges, implementation barriers, and mitigation strategies. Through retrospective analysis of survey responses, regulatory deliverables, and stakeholder feedback, we extracted and synthesized core principles into a generalizable regulatory-aligned innovation framework. Conclusions: The INTERCEPT regulatory-aligned framework aims to enhance digital health innovation by leveraging systematic regulatory integration to support project development and enable faster, more reliable market access for Software as Medical Device solutions, and contributes to regulatory science theory by establishing systematic principles for proactive compliance integration and offers practical guidance for digital health innovators navigating the complex EU regulatory landscape. Future research should validate the framework across diverse therapeutic areas and assess long-term impact on market success rates.

  • Background: Healthcare service quality is inherently multidimensional, yet document-level text analysis methods such as Latent Dirichlet Allocation (LDA) force patient reviews into single dominant topics. This simplification may systematically discard evaluative information when patients discuss multiple service dimensions with varying sentiments within the same review. Objective: This study compared document-level topic modeling (LDA) with GPT-based aspect-level sentiment analysis (ABSA) to address three research questions: (1) How much information is lost when collapsing multi-aspect reviews to single topics? (2) How prevalent are mixed-sentiment reviews, and what quality tensions do they reveal—both cross-aspect trade-offs and within-aspect ambivalence? (3) Do positive and negative reviews exhibit different structural patterns in aspect co-occurrence? Methods: We analyzed 2024 Google Reviews from 24 medical centers in Taiwan. Both LDA (K=7 topics) and GPT-based ABSA were applied to the same 5,467 reviews, ensuring fair comparison on identical data. The ABSA design employed structured prompts to extract aspects from seven predefined quality dimensions. Quality validation achieved Cohen κ=.82 against human annotation. Mixed-sentiment reviews were identified as those containing both positive and negative aspect evaluations, and cross-polarity couplings were analyzed to identify recurring trade-off patterns. Rating-stratified network analysis compared aspect co-occurrence patterns between positive reviews and negative reviews using Jaccard similarity. Results: Reviews discussed an average of 2.05 distinct aspects (SD=0.97), producing 51.2% information loss under LDA's single-topic assignment. Among multi-aspect reviews, 11.0% exhibited cross-aspect mixed sentiment, with Technical–Functional Divergence—praising Professional Quality while criticizing functional dimensions—appearing in 49.9% of these mixed-sentiment cases. Network analysis revealed differential bundling: operational dimensions co-occurred more strongly in negative reviews, whereas clinical dimensions co-occurred more strongly in positive reviews. Conclusions: Document-level topic modeling discards more than half of the evaluative information patients provide. Our findings reveal that patients cognitively decouple clinical competence from service delivery—Technical–Functional Divergence appeared in half of mixed-sentiment cases—and that positive and negative reviews organize quality dimensions differently. We recommend a complementary approach: topic modeling for exploratory discovery and ABSA for diagnostic assessment. For healthcare quality improvement, hospitals should separate clinical signals from operational signals in feedback dashboards.

  • Background: Continuous renal replacement therapy (CRRT) is a life-sustaining critical care intervention widely used for hemodynamically unstable individuals with acute kidney injury. Recent efforts, including standardized procedures, structured documentation, and quality monitory, have shown small improvements in CRRT delivery and safety. However, fragmented workflows and paper-based documentation limit the sustainable implementation of these improvements in routine practice. Objective: This study aimed to design and evaluate a CRRT information system to support standardized procedures, structured documentation, and quality monitory. Methods: A user-centered design approach, informed by Design Science Research (DSR) methodology, guided a multi-step process of identifying problems, defining objectives, and designing and evaluating the information system. The approach to design involved close collaboration within a nurse-led, 10-member multidisciplinary team comprising nephrologists, nurses, information technology specialists, and information engineers. Evaluation included six months of real-world clinical use with ongoing feedback collected through a dedicated WeChat workgroup and a System Usability Scale (SUS) survey of 27 CRRT care team members. Results: A role-based CRRT information system was developed, comprising 14 clinical modules and 6 core functions. The system embedded a continuous data-processing pipeline that enabled automated capture of treatment-related data directly from CRRT machines, creation of structured nursing documentation, and generation of quality indicators from structured data. During demonstration, workflow refinements—including dual-nurse verification and enhanced device data transmission—were incorporated following pilot testing. Over six months of clinical use, 42 user-reported issues were identified across three domains: data retrieval and calculation, fidelity of automatically generated clinical documentation, and interface appearance. Quantitative usability survey (n=27) demonstrated excellent usability (mean SUS score 95.19, SD 5.09). Conclusions: A CRRT information system integrating standardized clinical procedures, structured documentation, and ongoing quality monitoring supported complex clinical practice and management beyond simple digitization. Workflow-aligned, data-flow–enabled design may help future critical care information systems better support clinicians working in information-intensive environments.

  • Digital Health Engagement Behaviors in U.S. Family Caregivers: Trend Analysis Using HINTS 2019-2022

    Date Submitted: Mar 6, 2026
    Open Peer Review Period: Mar 9, 2026 - May 4, 2026

    Background: Digital health has provided caregivers with access to supportive resources without space-time restrictions. Caregivers’ digital health engagement behaviors can help them track their own health and that of care recipients as well as communicate with others. While digital health tools have become more prevalent since COVID-19, the trend of caregiver engagement has been less explored. Objective: This study examined the trends and factors associated with selected digital health engagement behaviors in family caregivers in the United States (U.S.), using the Health Information National Trends Survey (HINTS) datasets collected in 2019, 2020, and 2022. Methods: Our cross-sectional data analysis included 1,676 family caregivers. Dependent variables were: 1) access to online medical records (caregiver’s, care recipient’s); and 2) health-related use of social media (sharing health information, interacting with others, watching health-related videos). Independent variables were survey year, demographic, socioeconomic, caregiving, and Internet technology factors. Weighted multivariable logistic regression analyses were conducted. Results: Among 1,676 caregivers (2019: n = 570; 2020: n = 412; 2022: n = 694), access to online medical records increased from 2019 to 2022. Access to caregivers’ own records rose from 48.7% to 72.6% (P<.001), and access to care recipients’ records increased from 30.8% to 44.5% (P<.001). Health-related social media use also increased, including sharing health information (22.5% vs 39.1%, P<.001), interacting with others (16.6% vs 27.0%, P<.001), and watching health-related videos (49.5% vs 60.9%, P=.005). In adjusted analyses, higher education (college graduate vs ≤high school: OR = 2.75, 95% CI 1.56–4.85, P<.001) and having health insurance (OR = 2.40, 95% CI 1.24–4.68, P=.010) were associated with access to caregivers’ records. Female sex (OR = 1.96, 95% CI 1.36–2.84, P<.001) and spousal caregiving (OR = 2.14, 95% CI 1.26–3.65, P=.005) were associated with access to care recipients’ records. High-speed internet access was strongly associated with digital engagement outcomes (e.g., sharing health information: OR = 3.98, 95% CI 2.15–7.35, P<.001). Conclusions: Digital health engagement—including access to online medical records and the use of social media for health-related purposes—among U.S. family caregivers increased following the COVID-19 pandemic. These findings suggest that healthcare professionals and researchers should consider multifaceted factors, such as age, race/ethnicity, geography, education, insurance coverage, and digital access, when designing and implementing digital health tools and technology-based interventions. Future research should evaluate how digital technologies, automation of systems “talking” to other systems, including artificial intelligence (AI) can better support caregivers’ health information needs and care coordination. The varying learning curves for individuals and groups could further be explored for effective and efficient adoption and utilization.

  • Self-Esteem and Problematic Digital Use in Youth: The Role of Affective Symptoms and Objective Smartphone Behaviors in a Cross-Sectional Study

    Date Submitted: Mar 6, 2026
    Open Peer Review Period: Mar 9, 2026 - May 4, 2026

    Background: Problematic digital use among youth is associated with mental health concerns, yet the affective and behavioral mechanisms linking self-esteem to problematic digital use remain insufficiently characterized. Objective: To examine whether depressive and anxiety symptoms and objectively measured smartphone behaviors are associated with the relationship between self-esteem and problematic digital use among adolescents and young adults. Methods: This cross-sectional observational study was conducted between April 2022 and January 2023 in academic institutions in Grenoble, France. Participants were 171 adolescents and young adults aged 11 to 25 years using Android smartphones who completed self-report questionnaires alongside passive smartphone monitoring. Measures included self-esteem, depressive symptoms, anxiety symptoms, recreational smartphone time, delay to first connection in the morning, and nighttime digital disconnection (digital sleep). Problematic digital use was modeled as a latent construct encompassing excessive use, emotional regulation, and reactivity and assessed using a validated self-report scale. Results: Among 171 participants (meanage, 17.6 years, SD 3.0; 57% female), depressive symptoms mediated the association between self-esteem and problematic digital use (β = −0.33; 95% CI −0.45 to −0.21), with larger indirect effects than anxiety symptoms (β = −0.13; 95% CI −0.22 to −0.04). Recreational smartphone time was positively associated with problematic digital use (β = 0.28), whereas digital sleep was independently associated with lower problematic digital use (β = −0.24). Conclusions: Lower self-esteem was indirectly associated with problematic digital use primarily through depressive symptoms, which showed stronger associations than anxiety symptoms. Objective smartphone behaviors were independently associated with problematic digital use. Clinical Trial: This trial was registered at ClinicalTrials.gov (NCT07293208).

  • Social Determinants of Digital Health Intervention Use in Mainland China: A National Cross-Sectional Study

    Date Submitted: Mar 6, 2026
    Open Peer Review Period: Mar 9, 2026 - May 4, 2026

    Background: : Digital health interventions (DHIs), including telemedicine and artificial intelligence–enabled health tools, are increasingly integrated into health care systems worldwide. While these technologies have the potential to improve access and efficiency, unequal access to digital resources and health capabilities may create disparities in their use. Evidence on population-level determinants of digital health use remains limited in rapidly digitalizing health systems such as China. Objective: This study aimed to examine social and structural determinants of DHI use among adults in mainland China using the World Health Organization’s Social Determinants of Health (SDoH) framework. Methods: This cross-sectional study analyzed data from a nationally representative survey conducted in mainland China in 2024 among adults aged ≥18 years. The primary outcome was self-reported ever use of digital health interventions, including telemedicine, digital health applications, and AI-enabled health tools. Explanatory variables were categorized into five SDoH domains: economic stability, education and health-related capabilities, health care access and quality, neighborhood and built environment, and social and community context. Multivariable logistic regression models were used to estimate adjusted odds ratios (aORs) and 95% confidence intervals (CIs) for associations between social determinants and DHI use. Results: Among 34,672 participants, 14,565 (42.0%) reported ever using a DHI. Higher household income (≥6001 CNY vs ≤3000 CNY: aOR, 1.37; 95% CI, 1.29–1.46), higher educational attainment (bachelor’s degree or above vs junior high school or below: aOR, 1.49; 95% CI, 1.38–1.61), higher health literacy (per SD increase: aOR, 1.10; 95% CI, 1.07–1.13), and higher eHealth literacy (per SD increase: aOR, 1.20; 95% CI, 1.17–1.24) were associated with greater odds of DHI use. Health insurance coverage was associated with higher DHI use (aOR, 1.22; 95% CI, 1.11–1.34), whereas individuals aware of but not enrolled in family doctor services had lower odds (aOR, 0.65; 95% CI, 0.60–0.70). Difficulty paying medical expenses was associated with higher DHI use (aOR, 1.31; 95% CI, 1.22–1.41), while rural residence was associated with lower odds (aOR, 0.94; 95% CI, 0.89–1.00). Conclusions: DHI use in China is strongly associated with socioeconomic resources, health-related capabilities, and access to health care. These findings highlight the importance of addressing structural and social determinants to promote equitable adoption of digital health technologies in rapidly digitalizing health systems. Clinical Trial: NA

  • Background: Automated systems for detecting adverse drug reactions (ADRs) are increasingly common and carry high expectations from policymakers, researchers, healthcare professionals, and patients, yet evidence of their effectiveness and safety remains limited Objective: The aim of this systematic review was to identify the ethical, legal, organizational, social, and environmental implications of these systems. Methods: We conducted a systematic using the VALIDATE framework, we conducted a three-step approach: (1) defining scope through literature review and stakeholder consultation; (2) systematic review; (3) environmental inquiries. Results: Stakeholders prioritized research on feasibility, barriers, facilitators, alarm management, staged implementation, confidentiality, cybersecurity, and bias detection. The systematic review of ten studies revealed that leveraging new data sources and developing privacy-protection technologies is essential for upholding ethical and legal standards. Cybersecurity risks could expose patient information to unauthorized parties, while biases in training datasets can compromise fairness. Integrating ADR detection into clinical workflows and medication management systems can improve resource optimization and reporting rates. Establishing a positive reporting culture, supported by education and training for healthcare teams, is crucial to enhance ADR reporting. Conclusions: Careful planning is critical when implementing an early ADR detection system. Incorporating co-design methodologies can help align these automated systems with stakeholder needs and improve medication safety. Clinical Trial: Not requiered

  • Background: Bangladeshi adolescents, who constitute a fifth of the country's population, experience barriers in accessing sexual and reproductive health (SRH) information. Previous studies have shown that mobile health (mHealth) interventions provide adolescents with timely access to evidence-based curricula, gamified, and interactive content, sessions, and information. The widespread adoption of mHealth technologies among adolescents and their willingness to embrace emerging technologies are encouraging specialists to employ mHealth approaches to share health information. Despite the high mobile phone usage among adolescents in Bangladesh, there are a few mHealth interventions specifically targeting their SRH needs. Objective: We aimed to assess changes in SRH knowledge and awareness among adolescents in Bangladesh following exposure to "Mukhorito", an interactive mobile app-based intervention. Methods: This pilot study employing a pre-post non-randomized experimental approach was conducted in three selected secondary schools in Feni, Bangladesh, from June 2023 to March 2024. 46 students from class 9 across the three schools were recruited, with a minimum of 10 per school. Bivariate analyses were performed to assess the association between SRH knowledge and awareness scores with other covariates. Significantly associated covariates for both scores were used in building the adjusted linear regression models. Results: The adjusted models indicated a significant improvement in the end-line group compared with the baseline group for both knowledge (1.2 units; 95% CI: 0.8-1.6 units) and awareness scores (1.0 units; 95% CI: 0.3-1.5 units), indicating a high level of intervention effect. Conclusions: These findings demonstrate the potential of mobile app-based innovations to improve adolescent SRH education within a national program in resource-constrained settings, specially where conventional methods may be less effective.

  • Remotely Delivered Yoga Interventions for Pain Management: A Scoping Review

    Date Submitted: Mar 5, 2026
    Open Peer Review Period: Mar 5, 2026 - Apr 30, 2026

    Background: An expanding body of evidence suggests that yoga may be beneficial for pain management across a range of conditions. At the same time, healthcare delivery has evolved rapidly with the growth of telehealth, including advances in the remote delivery of yoga interventions. However, the literature lacks a comprehensive synthesis focused specifically on remote yoga interventions for pain. Objective: This scoping review aimed to map and characterize the extent and type of the existing evidence on remotely delivered yoga interventions for pain management. Specifically, we (i) examined general study characteristics, participant populations and intervention features, (ii) summarized reported findings on feasibility, safety and effects on pain-related outcomes, and (iii) identified research gaps to inform future investigation and practice. Methods: A systematic search was conducted in August 2025 across six databases to identify primary studies. Studies, including study protocols, conference abstracts and trial registrations, were eligible for inclusion if they reported primary data, across any study design, involving participants experiencing any type of pain, undergoing remote delivered yoga interventions (referred to also as online, virtual or tele-yoga), from any contextual setting. Eligibility was assessed through abstract and title screening and a subsequent full-text review independently by two reviewers. Results: A total of 82 sources of evidence were included, comprising 47 peer-reviewed publications, 1 preprint, 17 conference abstracts, and 17 trial registrations. Detailed data charting was conducted for peer-reviewed publications only. Overall, 3.199 participants were represented. Fewer than half of the studies examined pain as the primary condition, while the remainder assessed pain as secondary to other medical conditions or within non-clinical populations. Interventions varied considerably in duration, frequency, delivery format, and yoga style. Synchronous delivery was most common, Hatha yoga and adaptions were the most frequent style and eight- or twelve-week programs delivered twice weekly predominated. Feasibility was generally favorable, safety findings suggested a low risk of adverse events, and adherence was typically moderate to high, however reporting across these domains was inconsistent. Given the substantial methodological heterogeneity, conclusions about efficacy are limited, however reported findings indicate potential benefits for pain-related outcomes. Conclusions: Based on the current existing evidence, remotely delivered yoga for pain appears feasible and low risk, with signals of potential benefit. Future systematic reviews with formal quality appraisal and quantitative synthesis are needed to clarify effect sizes and the certainty of the evidence. Important gaps remain, including inconsistent reporting, limited comparative research on delivery formats, further investigation of intervention characteristics, underrepresentation of certain pain conditions and low- and middle-income settings.

  • Background: People living with advanced cancer experience more frequent and severe symptoms than people living with early-stage disease. Four common and distressing symptoms include sleep difficulties, worry-anxiety, fatigue, and depression. Cognitive-behavioral therapy (CBT) and acceptance and commitment therapy (ACT) interventions are effective for managing these symptoms but are often too time-intensive for people with multiple appointments, limited energy, and competing priorities. Brief, mobile health (mHealth) interventions provide an accessible alternative, particularly for those in rural communities with limited access to palliative and/or psychosocial oncology services. Objective: Based on our successful in-person/DVD-based pilot trial of a four session, integrated CBT-ACT symptom management intervention for advanced cancer patients, Finding Our Center Under Stress (FOCUS), this study tests the feasibility and acceptability of a mHealth translation of this intervention. Methods: In this single-group, feasibility trial, 11 people with advanced cancer were recruited through hospital-based oncology clinics representing four cancer types (breast, melanoma, multiple myeloma, prostate). Patients completed sociodemographic questions, initial patient-reported outcomes including sleep (ISI), anxiety (GAD-7, PSWQ), fatigue (FSI), and depression (CES-D) and a 7-day sleep diary via the mobile app. They then completed four modules focused on the self-management of sleep difficulties, worry-anxiety, fatigue, and depression. To assess feasibility, we examined recruitment, retention, and module completion. At the end of six weeks, to assess acceptability, participants completed the Internet Evaluation and Utility Scale and some participants completed a qualitative interview assessing their experience with the FOCUS app. We present quantitative and qualitative results as well as lessons learned in designing the application for this patient population. Results: Sixty-five percent entered the trial (N =11) and seventy percent completed more than half of the app. These participants gave strong ratings for FOCUS ease of use (3/4), convenience (3.7/4), utility (3.3/4), and ease of understanding (3.83/4). All participants (10/10) said they would recommend the app to other people with cancer and would return to the app with future problems. Participants’ favorite components were video recordings of other patients and the sleep and worry/uncertainty modules. Areas for improvement based on participant feedback included video quality for some components (i.e., lighting, sound), sleep diary ease of use, and a desire for professional guidance. Conclusions: The FOCUS intervention was successfully delivered via mobile technology and was feasible and acceptable per beta testing. The FOCUS mHealth app provides an evidence-based, accessible symptom management intervention for people with advanced cancer in rural communities. In accordance with participant feedback, for FOCUS 2.0 we will enhance video segments, incorporate a telehealth component to support app usage, and further develop the interactive and motivational features of the app. Future research will explore the effectiveness of this mHealth symptom management application via a randomized controlled trial.

  • Background: The rapid growth of digital technologies has generated large volumes of free-text data across healthcare, public health, and social research. These contain contextualised accounts of lived experience that are often absent from quantitative measures. Despite their value, these data remain underused because qualitative analysis is traditionally designed for in-depth work on smaller numbers. Computational methods, including topic modelling and large language models, are increasingly promoted as efficient solutions. However, concerns persist regarding interpretability, bias, hallucinations, and loss of contextual depth. Critically, there is no established human-centred framework for evaluating the quality of machine-generated outputs for qualitative analysis. Objective: 1) To develop an AI evaluation framework for assessing machine-generated outputs, 2) Evaluate different AI approached to textual data analysis Methods: We developed and applied a human-centred evaluation framework, GRACE (Grounded Review and Assessment of Computational Evidence), to assess the quality of machine-generated textual outputs. GRACE was derived from established qualitative appraisal tools and operationalised four core indicators: interpretability, actionability, nuance, and redundancy, using structured scoring and reflexive consensus. We compared classic probabilistic topic modelling (LDA), a deep learning embedding-based approach (BERTopic), and three large language models (LLMs: LLaMA-3, Copilot, DeepSeek), used alone or in combination with prior structural topic modelling (STM). These were applied to the same corpus (n = 1,044 free-text responses). LLM prompting was iteratively refined, with a single-shot STM-based configuration selected for final evaluation due to reduced hallucinations. All outputs were analysed using Machine-Assisted Topic Analysis. A rapid manual thematic analysis of a 15% subsample (n = 152) served as a pragmatic comparator. Results: Model outputs were variable, with different AI methods producing different results from the same dataset. GRACE evaluation indicated that LDA achieved the highest overall mean score (2.6/5), followed by BERTopic and topic modelling plus Copilot (2.5), topic modelling plus LLaMA-3 (2.2), and topic modelling plus DeepSeek (1.9). LDA generated broader conceptual patterns requiring interpretive refinement; while BERTopic produced narrower, more descriptive clusters with thematic overlap. LLM-only outputs were very poor. The combination of topic modelling and LLMs performed slightly better: the outputs were well structured but often superficial and repetitive. Conclusions: Computational models produced different interpretations of the same dataset, and performance did not align with technical complexity. Large language models were not suitable for thematic analysis, especially when applied to raw data, generating generalised and sometimes inaccurate outputs. Classical probabilistic modelling, particularly topic modelling + qualitative human analysis using the Machine Assisted Topic Analysis (MATA) approach provided the highest quality results. We argue that the key issue is not whether a model “works,” but whether it support meaningful, contextually grounded results. GRACE offers a simple, human-centred framework to support this assessment and build evidence base for analysis of free-text data that is useful and nuanced.

  • Pallvi–Family Focused Telepalliative Care: Development of a Complex Intervention

    Date Submitted: Mar 4, 2026
    Open Peer Review Period: Mar 5, 2026 - Apr 30, 2026

    Background: Telepalliative care, the use of telehealth in palliative care, has emerged as a strategy to improve access to specialist palliative services amid growing demand, workforce shortages, and increasing digitalization of health care. Although telepalliative care has demonstrated positive outcomes for patients, families, and clinicians, its integration into standard services remains inconsistent. Existing initiatives are often operationally focused and rarely grounded in programme theory or developed collaboratively with key stakeholders, limiting sustainability and contextual alignment, particularly in Nordic health systems that emphasize home-based palliative care. Objective: This study aimed to develop a family focused model of telepalliative care for clinical practice through active involvement of key stakeholders. Methods: A co-design qualitative study grounded in interpretive description was conducted. The development followed the British Medical Research Council’s guidance for the development and evaluation of complex interventions and represents the development phase. Key stakeholders including patients, family representatives, specialized palliative care team members, community care nurses, general practitioners, voluntary representatives, IT consultants, managers, and researchers, were purposively recruited. Data were generated through four scientific workshops across two Danish sites, supplemented by participant observations of video consultations and a short questionnaire inspired by the Normalisation Measure Development (NoMAD) questionnaire. Data were analyzed using abductive thematic analysis, with qualitative and quantitative findings converged and iteratively refined through stakeholder consensus. A programme theory and logic model guided development. Results: Eighteen stakeholders participated in the workshops, with additional input from clinicians through observations (6 consultations involving 22 participants) and questionnaires (n=10). Findings highlighted both alignment and tension between the proposed model and current clinical practice, particularly regarding when and for whom telepalliative care should be used, clinician digital competencies, and family involvement. These, and insights from previous studies, informed the primary output of the study which is Pallvi – Family Focused Telepalliative Care, a comprehensive, theory-informed model comprising of a structured consultation guide and two co-designed quick guides; one for health care professionals and one for patients and families. Pallvi integrates family focused care, shared decision-making, advance care planning, and the Calgary-Cambridge Communication Guide, operationalized across seven consultation phases. Conclusions: Through systematic stakeholder involvement and theory-driven development, this study produced a contextually and culturally aligned family focused model of telepalliative care. Pallvi addresses identified gaps in telepalliative care research by providing a structured, practical guide designed to support communication, family involvement, and cross-sectoral collaboration. Future research will focus on feasibility and implementation testing to assess acceptability, fidelity, and sustainability in clinical practice and implementation.

  • Promises and Challenges of the Use of Digital Interventions in Psychological Skills Training: A Systematic Review

    Date Submitted: Mar 3, 2026
    Open Peer Review Period: Mar 4, 2026 - Apr 29, 2026

    Background: Psychological skills training (PST) is a core component of sport psychology, supporting athletes’ performance, well-being, and capacity to manage competitive stress. However, access to high-quality, practitioner-led PST is often constrained by time, cost, availability of trained professionals, and stigma surrounding help-seeking. In response, digital interventions such as mobile applications, biofeedback systems, and immersive technologies have been increasingly adopted to deliver PST in more scalable and flexible formats. Despite rapid growth in this area, evidence regarding the promises and challenges of digital PST remains fragmented across modalities and outcome domains. Objective: This systematic review synthesizes empirical evidence on the use of digital interventions for delivering PST in athlete populations. Specifically, it maps the digital modalities employed, the psychological skills and frameworks targeted, the populations and sporting contexts studied, and the promises and challenges reported in relation to effectiveness, feasibility, and implementation. Methods: We conducted a PRISMA-compliant systematic review of English-language studies published between 2000 and 2025. Three databases (Scopus, Web of Science Core Collection, and ProQuest Dissertations and Theses) were systematically searched, and additional records were identified through a manual search. Eligible studies examined digital or technology-based interventions deployed to support PST outcomes in athlete populations and reported empirical quantitative, qualitative, or mixed-methods findings. Two reviewers independently screened records and extracted data, resolving discrepancies through discussion. Results: Thirty-six studies met the inclusion criteria, encompassing virtual reality-based interventions, mobile applications, and biofeedback or neurofeedback systems. Across modalities, digital PST interventions targeted a range of psychological skills, including stress and anxiety regulation, attentional control, imagery ability, self-talk, and emotional regulation. Reported promises included improvements in affective, cognitive, physiological, and performance-related outcomes, enhanced accessibility, flexibility, and engagement of PST delivery, and potential for skill transfer beyond sport. However, recurring challenges were also identified, such as limited personalization, variable user engagement, technical and cost barriers, and inconsistent or weaker efficacy relative to traditional PST methods. Conclusions: Digital interventions offer a meaningful extension to traditional PST by widening access, enhancing immersion, and providing real-time feedback that supports psychological skill development. However, their effectiveness appears constrained by methodological variability, limited personalization, and implementation challenges. Future research should prioritize rigorous longitudinal designs, clearer alignment with PST theory, and hybrid delivery models in which digital tools complement practitioner expertise, to ensure digital PST enhances rather than dilutes psychological practice.

  • Digital Frailty in Ageing Societies: Introducing a Digital Health Vulnerability Index

    Date Submitted: Mar 3, 2026
    Open Peer Review Period: Mar 3, 2026 - Apr 28, 2026

    Digital health is now embedded in routine care through patient portals, teleconsultations, remote monitoring, digital triage, and other hybrid service models. While these changes can improve access and efficiency, they may also create new barriers for older adults who have limited cognitive, sensory, functional, or social capacity to engage with digitally mediated care. Current constructs such as digital literacy, digital exclusion, and conventional frailty only partly explain this problem because they do not fully capture the mismatch between the digital demands of healthcare systems and the real world capabilities and supports available to patients. This Viewpoint introduces digital frailty as a clinically relevant, multidimensional state of vulnerability that arises when a person’s intrinsic capacity and available support are insufficient to meet the digital requirements of healthcare. We argue that digital frailty should be understood not as a synonym for age, disability, or low digital confidence, but as a relational and potentially modifiable mismatch between individuals and care environments. Framing the issue in this way shifts attention from blaming patients to designing safer and more equitable systems. To operationalize this concept, we propose a Digital Health Vulnerability Index as a pragmatic framework for identifying patients at risk of digitally mediated care failure. The framework focuses on four proximal domains of vulnerability, namely access, skills, confidence or trust, and support, and is paired with brief consideration of hearing, vision, and cognition to improve clinical interpretability. Rather than functioning as a static label, the index is intended as a routable mechanism to trigger proportionate responses such as assisted digital support, proxy enabled access, simplified workflows, and analogue alternatives for safety critical steps. We further propose proportionate universalism as the most appropriate implementation principle, so that digital support is universal in reach but calibrated in intensity according to need. This approach has implications beyond individual assessment and extends to pathway design, procurement, governance, reimbursement, and digital inclusion policy. In ageing societies, digital vulnerability should be recognized as a determinant of functional access to care. A digitally inclusive health system therefore requires not only better technology, but also better identification, adaptation, and accountability for the patients most at risk of being left behind.

  • The Role of Measurement in Identifying High-Intensity Secure Message Senders: An Observational Study

    Date Submitted: Feb 27, 2026
    Open Peer Review Period: Mar 3, 2026 - Apr 28, 2026

    Background: The growing volume of secure messaging within the patient portal has imposed significant demands on clinicians and contributed to burnout. Little is known about the characteristics of patients who comprise high-volume message senders, and we lack a nuanced understanding of patient messaging intensity beyond measures accounting for sheer volume. Objective: Our objective was to characterize older adult patients (65+) with high secure messaging volume, examining both patient characteristics and other aspects of their messaging intensity such as messaging frequency, length, and messaging use relative to patient portal logins and healthcare encounters. Methods: We analyzed electronic medical record (EMR) and patient portal data from a large academic health system, encompassing 16,023 older adults who sent 199,952 messages over a 12-month period. We developed five measures to account for secure messaging intensity. Our primary measure of messaging intensity was based on message volume; high-volume message senders were identified using outlier analysis based on patients’ aggregate number of messages sent during the observation period. Additional measures of messaging intensity included identifying individuals with concentrated periods of messaging, message length (character count), a ratio of messages to portal logins and a ratio of messages to healthcare encounters. We compared sociodemographic characteristics, health status, and messaging intensity of high-volume secure messaging senders to other message senders. We also identified patients who were classified as high-intensity message senders based on all five measures of messaging intensity (‘super-senders’). Results: Of 16,023 older adult patients who sent at least one message during the observation period, 1,298 (8.1%) were classified as high-volume message senders; these patients accounted for 39.7% of total messages. High-volume message senders, compared to all other message senders, were more likely to be White (80.4% vs. 72.5%, p < 0.001), have higher comorbidity scores (2.6 vs. 1.8, p <0.001), and higher incidence of cancer (35.8% vs. 22.8%, p<0.001) and dementia (8.3% vs. 6.1%, p < 0.002). High-volume message senders were also more likely to be identified as having concentrated periods of messaging, to send longer messages, and to send more messages in relation to patient portal logins and healthcare encounters. A small subgroup of patients classified as high-volume senders were also classified as high-intensity across all four of the other measures of messaging intensity (59/1,298; 4.5%), the ‘super senders’. Conclusions: High-volume message senders represent a small but distinct group of older patients who send a disproportionate share of messages to clinicians. Triangulating multiple measures of messaging intensity can help provide additional context about patient messaging behavior and help to identify patients that may most benefit from targeted outreach while potentially easing clinicians' inbox workload.

  • Predicting Healthcare Professionals’ Use of Telehealth in China: A Cross-Sectional Study

    Date Submitted: Mar 1, 2026
    Open Peer Review Period: Mar 3, 2026 - Apr 28, 2026

    Background: While telehealth has become a transformative tool enhancing healthcare accessibility and efficiency, adoption rates in China remain low. Chinese healthcare professionals’ low telehealth adoption rates are poorly understood. Objective: Our study investigates the key factors influencing Chinese healthcare professionals’ intention to adopt and actual use of telehealth. Based on the results from estimating an integrated telehealth use framework, we also make recommendations for improving healthcare professionals’ telehealth adoption. Methods: Data on 10,372 healthcare professionals from the 2023 Xi’an Healthcare Worker Survey were analyzed, utilizing descriptive statistics (chi-square test, group differences), reliability testing (Cronbach’s α coefficients), Discriminant validity analysis (square root of average variance extracted) and fit tests. Based on our integrated telehealth use framework, structural equation modeling was employed to test hypotheses and path relationships, including multi-group analysis to examine demographic moderating effects. Results: Confirming our hypotheses on telehealth intention to use and actual use, the structural equation model showed strong fit indices. Key predictors of behavioral intention to use telehealth included effort expectancy, price value, performance expectancy, and social influence. Behavioral intention and facilitating conditions positively influenced actual use behavior, while demographic characteristics moderated specific relationships. Conclusions: Our study identifies critical factors influencing healthcare professionals’ adoption of telehealth, including performance expectancy, social influence, and facilitating conditions. It offers an integrated framework to assess behavioral intentions and provides practical insights for advancing telehealth implementation in China. Tailored strategies for diverse demographics and institutions are essential for promoting sustainable adoption. Clinical Trial: This study was reviewed and approved by the Biomedical Ethics Committee of Xi’an Jiaotong University (approval number: XJTUAE2646).

  • Background: The COVID-19 pandemic has spurred an unprecedented collection of multi-omic and clinical data of patient cohorts. However, fragmented datasets, inconsistent terminology, and poor integration with existing knowledge of molecular pathways hinder effective analysis. Objective: The objective of our development and usability study was to develop SPOKE-C19, a web-based knowledge graph tool. Methods: We designed this tool to integrate multi-omics and clinical data from the INCOV and RECOVER cohorts with SPOKE, a knowledge graph with over 27 million nodes of biomedical entities linked by 53 million relationship edges. The platform allows mapping of empirical data to the network of known mechanistic relationships, facilitating interpretation of novel associations observed in post-acute sequelae of SARS-CoV-2 infection (PASC) studies. Results: SPOKE-C19 offers an intuitive user interface for researchers to query biomedical variables and to link empirical associations observed in clinical or experimental studies to generic mechanistic pathways and to precomputed relationships extracted from large COVID studies. Conclusions: The comprehensively documented interactive COVID-C19 platform promotes “connecting the dots” between disjoint domains and encourages cross-disciplinary collaboration, which is particularly important given the multi-faceted nature of PASC. Ultimately, we designed the application as a comprehensive platform for researchers, offering robust tools that support innovative discoveries and contribute to a deeper understanding of PASC and its implications.

  • Surgeons’ Perceptions of Machine-learning‑based Prognostic Tools for Spine Surgery: A Qualitative Study

    Date Submitted: Feb 20, 2026
    Open Peer Review Period: Feb 21, 2026 - Apr 18, 2026

    Background: Machine-learning-enabled prognostic models are increasingly proposed to support surgical decision-making for degenerative lumbar disorders, yet their clinical adoption remains limited. Understanding how surgeons perceive these tools is critical for effective implementation, particularly given the high-stakes nature of spine surgery where decisions carry long-term functional consequences and medico-legal implications. Objective: This study aimed to explore how consultant-level spine surgeons perceive machine-learning-based prognostic tools, including their trustworthiness, clinical utility, usability, workflow integration, and implications for patient counselling and shared decision-making. Methods: A qualitative study using semi-structured one-to-one interviews was conducted with 11 consultant-level orthopaedic and neurosurgeons practising in Singapore (response rate: 73% of 15 invited). Participants had a mean of 10.9 years (range: 8-25 years) of spine surgery experience; 82% (n=9) were male. Interviews (range: 15-57 minutes) were transcribed verbatim and analysed using Braun and Clarke's six-step reflexive thematic analysis. Data collection continued until information power was achieved, with three additional interviews completed after thematic sufficiency was reached as participants had already consented. The study was designed and reported in accordance with COREQ criteria. Results: Three overarching themes were developed. Trust contingent on data integrity revealed that surgeons' confidence depended fundamentally on data quality, local representativeness, labelling credibility, and rigorous validation, with participants consistently emphasising population representativeness as a trust prerequisite. Surgeons operationalised trust through accessible performance metrics, with most identifying AUROC as their preferred credibility heuristic. Pragmatic orientation as important for implementation demonstrated that usability and seamless electronic health record integration were non-negotiable prerequisites, with surgeons explicitly stating that manual data entry would preclude adoption. Medico-legal concerns were prominent, with participants emphasising that decision responsibility remains with the clinician. Surgical decision-making as a delicate dance between art and science reflected how prognostic outputs were positioned as adjunctive inputs to be reconciled with experiential judgement and patient heterogeneity. Surgeons emphasised that identical prognostic information would be interpreted differently based on practice philosophy, and most highlighted the inherent divergence between technical success and patient-perceived success, underscoring prognostic tools' value for expectation management rather than deterministic prediction. Conclusions: This first qualitative study of spine surgeons' perceptions reveals that adoption of machine-learning-based prognostic tools is contingent on data integrity, pragmatic workflow integration, and alignment with professional judgement and not predictive performance alone. Surgeons expressed cautious openness, viewing these tools as valuable in complex cases for clarifying outcome expectations without displacing clinical responsibility. Meaningful implementation requires robust data governance, contextually grounded validation, seamless electronic integration, and explicit positioning of machine learning as supportive rather than substitutive of surgical judgement. These findings provide empirically grounded guidance for developing clinically acceptable and implementation-ready prognostic decision support systems.

  • Background: Diabetes mellitus affects approximately 537 million adults globally, with projections indicating an increase to 643 million by 2030. Mobile health applications (mHealth apps) offer promising support for diabetes self-management, yet adoption rates remain low. Understanding the factors influencing patients' intentions to use mHealth apps is essential for designing effective interventions. Objective: To develop and empirically validate an extended Unified Theory of Acceptance and Use of Technology (UTAUT) model incorporating personal innovativeness and attitude to explain behavioral intention to use mHealth apps for diabetes management. Methods: A cross-sectional survey was conducted with 485 Chinese adults. The measurement and structural models were assessed using Partial Least Squares Structural Equation Modeling (PLS_SEM). Results: Performance expectancy (β = .110, t = 3.401, P < .001), effort expectancy (β = .226, t = 5.942, P < .001), social influence (β = .112, t = 2.953, P =.002), facilitating conditions (β= .095, t = 2.476, P =.007), and personal innovativeness (β = .365, t = 9.280, P < .001) significantly influenced attitudes toward mHealth apps. Performance expectancy (β = .069, t = 2.239, P =.01), effort expectancy (β = .377, t = 8.939, P < .001), social influence (β = .123, t = 3.279, P < .001), and personal innovativeness (β = .116, t = 3.459, P < .001) significantly affected behavioral intention, while facilitating conditions did not (β = .041, t = 1.418, P =.07). Attitude significantly influenced behavioral intention (β = .337, t = 8.010, P < .001). Additionally, attitude significantly and positively mediated the relationships between performance expectancy (β = .037, t = 3.128, P < .001), effort expectancy (β = 0.076, t = 4.568, P < .001), social influence (β = .038, t = 2.775, P =.003), facilitating conditions (β = .032, t = 2.433, P =.007), and personal innovativeness (β = .123, t = 5.787, P < .001) and the behavioral intention to use mHealth apps for diabetes management. The model explained 31.7% of the variance in attitude and 51.5% in behavioral intention. Conclusions: The extended UTAUT model effectively explains mHealth app adoption for diabetes management by integrating personal innovativeness and attitude. Emphasizing app utility, usability, social influence, and fostering positive attitudes can enhance adoption. These insights inform healthcare providers and developers aiming to increase mHealth engagement among patients with diabetes.

  • Adaptation and Potential of Virtual Reality in Substance Use Disorders: A focus review

    Date Submitted: Feb 17, 2026
    Open Peer Review Period: Feb 19, 2026 - Apr 16, 2026

    Background: Substance use disorder (SUD) is a chronic, relapsing condition characterized by compulsive substance use and dysregulation in reward and control systems. Although effective pharmacological and psychosocial treatments are available, their impact is often limited by barriers such as stigma, poor adherence, and restricted access to care. Virtual Reality (VR) has emerged as a digital health intervention offering an adjunctive approach by providing immersive, interactive environments that may enhance engagement, simulate real-world triggers, and support therapeutic learning. Objective: This focus review aimed to map and synthesize the existing evidence for VR-based interventions in SUD treatment. We examine both therapeutic applications across established treatment frameworks and experimental approaches, identify key opportunities for future research and clinical innovation. Methods: We searched electronic databases including PubMed/MEDLINE, Science Direct and MDPI covering 2004 to 2025. Two reviewers independently screened for relevant studies and extracted study characteristics. Studies addressing VR applications for substance use disorders including peer-reviewed articles, randomized controlled trials, protocols and pilot studies published in English were selected. Any discrepancies were resolved through discussion. Results: A total of 26 studies or protocols were included in this review. Overall, the studies reviewed are broadly categorized into 6 sub-groups based on the type of the VR intervention and treatment class delivered. The reviewed literature indicates that VR-based cue exposure therapy is associated with reductions in craving and physiological reactivity for nicotine, alcohol, and cannabis use, with more limited and preliminary findings for opioid use disorder. VR relaxation and stress-management environments were linked to decreases in craving, stress, and pain among individuals with opioid and alcohol use disorders. VR-enhanced cognitive-behavioral interventions showed improvements in attention, cognitive flexibility, and emotion regulation. Motivational, social skills, and gamified VR interventions were associated with increased engagement, reduced stigma, enhanced self-efficacy, and improved treatment retention. Conclusions: This focus review contributes to the growing digital health literature by synthesizing current evidence on VR-based interventions for SUDs. The findings suggest that VR may serve as a flexible adjunct to existing treatments, with the potential to address persistent barriers to engagement and access. Further rigorously designed studies are needed to evaluate long-term effectiveness, optimize VR design, and support their integration into routine clinical practice.

  • Background: Mobile health (mHealth) app effectiveness may be limited by low engagement. Increasing understanding of factors influencing engagement may help. Paid mHealth app subscription and renewal are two metrics of particular interest to commercial app developers. Objective: Objective: To identify homogenous user subgroups (ie, behavioral phenotypes) within a paid mHealth app context and examine associations with app subscription and renewal. Methods: Methods: In this 6-month prospective cohort study, latent class analysis (LCA) was conducted with users of a paid mHealth app. Users completing a 7-day free trial between November 2023 and January 2024 were included. LCA produced phenotypes using survey responses (eg, chronic disease status), device-assessed health data (eg, daily step count), and 7-day free trial period engagement data (eg, number app opens). Odds ratios (ORs; P < .05) assessed associations between phenotypes and subscription/renewal. Results: Results: The sample included 934 users (mean age, 41.53 [SD, 9.65] years). Based on LCA fit indices five distinct phenotypes were formed: (1) Highly engaged subscribers, (2) Subscribers with multimorbidity, (3) Healthy subscribers, (4) Non-subscribers with multimorbidity, and (5) Healthy non-subscribers. Phenotypes 1–3 had greater odds of subscribing (OR = 21.31 [8.56, 53.06]; OR = 7.11 [4.04, 12.50]; OR = 8.28 [4.26, 16.08], respectively) than phenotype 4 (OR = 0.82 [0.48, 1.41]), compared to phenotype 5, the reference scenario. Additionally, renewal odds for phenotypes 1–4 were 1.06 [0.62, 1.81], 0.90 [0.54, 1.49], 0.99 [0.58, 1.69], and 0.93 [0.48, 1.80], respectively (vs. reference). Conclusions: Conclusions: Behavioral phenotypes associated with subscription likelihood were identified using data collected during the 7-day trial period. These phenotypes may be strategically targeted with future intervention to boost early engagement and long-term behavior change potential.

  • Background: It is increasingly common that patients are given the option to receive health information using digital technology. Augmented reality (AR) is an emerging technology which may enable patients to better appreciate anatomy pertinent to their disease process. Objective: Our objective was to understand patients’ perspectives about augmented reality in the context of shared decision making (SDM) for oncoplastic breast surgery. Methods: Three focus groups with a total of 17 participants without breast cancer were recruited from general surgery out-patient clinics in a university teaching hospital. Participants interacted with a Microsoft HoloLens 2™ head-mounted display presenting an anonymised three-dimensional holographic model of a breast cancer, which was used as a stimulus to prompt discussion. Anonymised interview audio transcripts were transcribed verbatim and analysed using thematic analysis. Results: Analysis revealed four themes: 1) Seeing as believing – AR enhanced participants ability to visualise abstract anatomy and aid understanding, 2) Being in the surgeon’s shoes – the technology offered insight into the clinicians perspective, with concerns raised about the emotional impact, 3) How technology influences trust – AR reinforced confidence in the shared decision-making process when introduced by trusted clinicians, 4) Involving people in my life – shared viewing with family or friends offered support in the decision-making process. Conclusions: Participants viewed AR as a promising tool to enhance knowledge and add value in the process of SDM beyond the reach of current information giving. They also expressed caution, emphasising the need for careful introduction and ongoing clinician support to ensure meaningful use. Their varied responses highlighted the challenge inherent to introduction of digital technologies such as AR – the question of how much, and what type of technology best supports patients without causing information overload. Trust in clinicians remained central to the perceived value of the technology. These findings highlight AR’s potential to enhance SDM when thoughtfully integrated into clinical practice.   Clinical Trial: N/a

  • Background: As oncology workflows integrate increasingly autonomous artificial intelligence (AI) agents, health systems face uncertainty regarding operational impacts. Traditional linear forecasting methods fail to capture second-order effects such as governance saturation, induced demand, and bottleneck migration. To navigate this complexity, the emerging field of Medical Futures Studies requires methodologies that bridge qualitative strategic foresight with quantitative operational modeling. These system-level dynamics directly influence patient access to timely diagnosis and treatment, with direct consequences for patient access, treatment delays, and health system resilience. Objective: To develop a proof-of-concept framework for stress-testing AI adoption strategies in oncology by coupling qualitative scenario planning with computational discrete-event simulation (DES). Methods: We defined a strategic state space using two orthogonal axes, AI automation intensity and data interoperability, resulting in four distinct futures scenarios. We translated these qualitative narratives into a quantitative DES model to simulate a 3-year operational horizon. The model quantified system performance (Referral-to-Treatment Interval [RTTI], throughput), volatility, and resource constraints across different adoption trajectories. Results: The scenario planning phase yielded four operational archetypes (analog oncology, automation islands, interconnected clinicians, and AI-orchestrated care) with distinct constraints, risks and failure modes. In the simulation, the fully integrated scenario maximized capacity (1,244 patients/year) and halved the mean RTTI to 14.9 days, a magnitude comparable to major pathway redesign interventions. Isolated automation without data infrastructure led to reduced system performance, increasing RTTI by 26% (37.1 days) and reducing throughput to 647 patients/year due to administrative governance saturation. The model demonstrated a structural bottleneck migration: successful upstream AI adoption shifted binding constraints from diagnostic scanners to downstream chemotherapy infusion units, while missing data interoperability resulted in governance constraints. Pathway optimization analysis indicated that a coordinated strategy prioritizing early improvements in data interoperability reduced transition volatility compared to an automation-first approach. Conclusions: Integrating qualitative scenario planning with quantitative simulations enabled a systematic evaluation of oncology AI adoption strategies. As a proof of concept, it offers a replicable framework for health leaders to model future scenarios of digital transformation in times of high uncertainty. Subsequent work should expand this methodology to incorporate financial and health equity dimensions, establishing simulation-based scenario planning as an important tool in Medical Futures Studies.

  • Impact of Digital Detox Awareness and Practices on Health Profession Students’ Mental Wellbeing, Physical Health, and Academic Performance: A Cross-Sectional Study

    Date Submitted: Feb 17, 2026
    Open Peer Review Period: Feb 18, 2026 - Apr 15, 2026

    Background: Health profession education students exhibit a higher rate of excessive digital technology use compared to their peers. Although the interaction of technology with student well-being has become more pronounced, the lack of awareness about digital detox among students in technology-intensive healthcare disciplines, along with the scarcity of studies exploring their practices, is concerning. Objective: This study aimed to investigate the patterns of social media usage and potential relationships between digital detox practices, mental well-being, physical health, and academic performance. Methods: A cross-sectional survey design was employed at King Saud bin Abdulaziz University for Health Sciences (KSAU-HS) in Riyadh. The sample consisted of 471 students from the health professions. Validated surveys were used, including the Social Media Disorder Scale, Digital Detoxification Awareness Questions, Kessler Psychological Distress Scale (K-6), and physical health assessments. The relationships between the study variables were analyzed using the chi-square test and ANOVA, with a significance level of 0.05. Results: A total of 471 students were included, with the majority being female (n = 291, 61.8%), single (n = 440, 93.4%), and aged between 18 and 37 years (M = 21.62, SD = 2.30). Participants reported an average daily social media usage of 7.07 ± 4.11 hours, with 31.6% of the sample classified as problematic users. Digital detox awareness was 59.7%, and 58.6% reported having experienced a digital detox. The most common strategies reported were avoiding phone use (69.1%) and muting notifications (70.3%). Participants reported eye strain (59.0%), neck pain (56.7%), and back pain (49.7%) due to the use of smartphones. Significant associations were found between social media use, gender, college affiliation, awareness of digital detox, level of physical activity, and sleep patterns (p < 0.005). A positive correlation was found between GPA and digital detoxification (p = 0.01). Social media use was significantly associated with the mental well-being of the participants (F = 214.096, p < 0.001) and with their academic performance (p = 0.04) Conclusions: The relationships between digital behavior, physical health, mental well-being, and academic performance of health profession students are complex and intertwined. The practice of digital detox, as observed, offers improvements in various aspects of students' lives; therefore, incorporating digital wellness strategies into the curriculum is vital for preparing students as professionals and enhancing student outcomes. Clinical Trial: NRR24/007/11

  • Reliability, Quality, and Content of YouTube Videos on Xerostomia: A Cross-Sectional Study

    Date Submitted: Feb 17, 2026
    Open Peer Review Period: Feb 18, 2026 - Apr 15, 2026

    Background: Xerostomia is a prevalent condition that negatively affects quality of life. Patients increasingly seek health-related information through online platforms such as YouTube. Given the growing role of social media in digital health communication, evaluating the reliability and quality of publicly accessible video content is essential. Objective: This study aimed to assess the reliability, quality, and content characteristics of YouTube videos related to xerostomia. Methods: In this cross-sectional study, a YouTube search was conducted on January 10, 2025, using the keyword “dry mouth.” The first 100 videos retrieved using the relevance filter were screened. After applying inclusion and exclusion criteria, 46 videos were included in the analysis. Video reliability was evaluated using the Modified DISCERN (mDISCERN) instrument, while quality was assessed using the Global Quality Score (GQS) and the Video Information and Quality Index (VIQI). Videos were further categorized as “useful” or “misleading”. Engagement metrics, including number of likes, views, comments, interaction index, and viewing rate, were recorded. Statistical analyses were performed using SPSS version 22.0, with significance set at P < .05. Results: A substantial proportion of videos demonstrated low reliability and quality. Approximately half of the included videos were classified as misleading. Useful videos had significantly higher mDISCERN, GQS, and VIQI scores compared with misleading videos (P < .05). In addition, useful videos showed significantly higher engagement metrics, including number of likes, views, comments, and viewing rate (P < .05). Positive correlations were observed between reliability and quality scores and engagement parameters. Conclusions: A considerable portion of YouTube videos on xerostomia contains low-quality or misleading information. Although higher-quality videos tend to receive greater user engagement, the presence of inaccurate content remains concerning. Increased involvement of healthcare professionals and academic institutions in producing evidence-based digital content may improve the quality of online health information. Clinical Trial: This cross-sectional study evaluated YouTube videos related to xerostomia. As the study analyzed publicly available data on an open-access platform and did not involve human participants or identifiable personal information, ethical approval was not required, consistent with previous similar studies.

  • Background: The growing elderly population can directly impact countries’ productivity and pose significant challenges for governments, becoming a potential public health concern due to the increasing prevalence of health issues such as frailty, dementia, mobility limitations, and cardiovascular diseases. One promising approach is to integrate emerging technologies, such as wearable devices, machine learning, and smart sensors, to support older adults in their daily activities. These technologies can promote independence by enabling the safe execution of essential tasks while allowing continuous, 24-hour monitoring through mobile health systems. Objective: This research paper aims to evaluate the current use of consumer-grade wearable technologies, in combination with machine learning techniques, to promote autonomy and enhance daily activities among older adults. Methods: We conducted a systematic review in accordance with the Cochrane Handbook and Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to synthesize evidence on the use of wearable technologies combined with artificial intelligence, particularly machine learning methods, to support daily living and prevent falls among older adults. The search was conducted in PubMed, MEDLINE, Scopus, the Institute of Electrical and Electronics Engineers (IEEE) Xplore, and the Association for Computing Machinery (ACM) Digital Library from their inception through April 2025, with no date or language restrictions. Results: Twenty-four studies were included. Mos studiest were observational or methodological and relied primarily on inertial sensing from wrist- or waist-mounted devices. The main application domains were activities of daily living monitoring, gait and mobility assessment, cognitive impairment, Parkinson’s disease symptoms, fall risk and detection, and frailty assessment. Classical machine learning models (e.g., support vector machines and random forests) and deep learning architectures (e.g., CNNs and LSTMs) were both widely used. However, studies were highly heterogeneous, frequently involved small samples, and rarely performed external validation or reported clinically actionable outcomes. Conclusions: Consumer-grade wearable devices, when combined with machine learning, show promise in supporting autonomy, daily activity monitoring, and fall-related safety in older adults. Nevertheless, the current evidence base is limited by methodological heterogeneity, small sample sizes, scarce external validation, and limited clinical integration. Future research should prioritize real-world evaluations, standardized reporting (e.g., TRIPOD-AI), interdisciplinary co-design, and patient-centered outcomes to enable translation into routine care. Clinical Trial: International Prospective Register of Systematic Reviews (PROSPERO) Registration: CRD420251044449

  • Re-assembling primary care therapies with virtual technologies: An Actor Network Theory Analysis of Pandemic Telerehabilitation

    Date Submitted: Feb 16, 2026
    Open Peer Review Period: Feb 17, 2026 - Apr 14, 2026

    Background: The COVID-19 pandemic triggered an abrupt transition to virtual rehabilitation across physiotherapy, occupational therapy, and respiratory therapy. While telerehabilitation research has documented feasibility and patient satisfaction, less is known about how professionals navigated the destabilization and reassembly of care practices during this transformation. Existing literature frames virtual care as a technical substitution for in-person services, overlooking the deeper reconfiguration of the socio-technical networks that organize therapeutic work. Objective: Applying actor-network theory (ANT), we examined how rehabilitation professionals reconfigured their practices through technology during the first year of the pandemic. We explored how digital tools, domestic spaces, and new sensory practices reshaped therapeutic presence, professional identity, and the environments in which care was enacted. Methods: We conducted a secondary analysis of longitudinal diary-interview data collected from 16 Canadian rehabilitation professionals (occupational therapists, physiotherapists, and respiratory therapists) working in community-based primary care in Ontario and Manitoba (2020-2021). Participants recorded audio diaries over 12 weeks and completed two follow-up interviews. Analysis followed an interpretive approach informed by Science and Technology Studies, tracing how human and technological actors were enrolled, adapted, and redefined within emerging care assemblages. Results: Three interconnected processes characterized the reconfiguration of rehabilitation: (1) technology as active participant, where digital platforms mediated rather than merely transmitted therapeutic reasoning and clinical decision-making; (2) reconfiguration of therapeutic presence, as sensory attention and embodiment were redistributed across screens, sounds, and new forms of spatial choreography; and (3) enrollment of domestic spaces as clinical environments, as clinicians' and patients' homes became sites of care shaped by new ethical, material, and relational dynamics. These processes reveal that virtual rehabilitation constituted a new form of care co-produced by humans, technologies, and spaces rather than a digitized replication of traditional practice. Conclusions: The pandemic exposed rehabilitation as a socio-technical practice sustained through the coordination of multiple actors rather than professional expertise alone. Virtual care redefined therapeutic presence when traditional boundaries between clinical and domestic, human and technological, were blurred. Recognizing virtual care as a distinct modality underscores the need to integrate technology-mediated competencies into rehabilitation education and practice. Future research should incorporate patient perspectives and direct observation to trace how these care networks evolve.

  • Background: Herpes zoster (HZ) imposes a substantial disease burden, yet vaccine uptake remains suboptimal in China. While eHealth literacy is a known determinant of health behaviors, its role in bridging socioeconomic disparities and its varying impact across different age groups of vaccine-eligible adults remain understudied. Specifically, it is unclear whether eHealth literacy acts as a "compensatory resource" for disadvantaged populations and if the digital skills required to reduce hesitancy differ between middle-aged and older adults. Objective: This study aimed to examine the association between eHealth literacy and HZ vaccine hesitancy among adults aged 40 years and older in Shanghai, China, with a specific focus on identifying age-dependent "digital thresholds" and the compensatory effect of literacy on socioeconomic status (SES). Methods: A community-based cross-sectional study was conducted from October to December 2022 across three districts in Shanghai. A total of 1302 adults aged ≥40 years were recruited via convenience sampling. eHealth literacy was assessed using the eHealth Literacy Scale (eHEALS). Multivariable logistic regression models were used to analyze the associations, adjusting for sociodemographic characteristics, health status, and behaviors. Stratified analyses were performed to evaluate interactions among literacy, age, and SES. Results: The prevalence of HZ vaccine hesitancy was 88.2% (1149/1302). In the fully adjusted model, participants with medium (odds ratio [OR] 0.538, 95% CI 0.326-0.886; P=.015) and high (OR 0.472, 95% CI 0.264-0.844; P=.011) eHealth literacy demonstrated significantly lower odds of hesitancy compared to those with low literacy. Age-stratified analyses revealed a distinct "digital threshold" effect: for middle-aged adults (40–59 years), medium literacy was sufficient to significantly reduce hesitancy (OR 0.501, 95% CI 0.265-0.949; P=.034), whereas older adults (≥60 years) required high literacy to achieve a significant protective effect (OR 0.347, 95% CI 0.136-0.882; P=.026). Crucially, eHealth literacy exhibited a strong compensatory effect for socioeconomic disadvantage. Among participants with low SES, high eHealth literacy was associated with an 83.1% reduction in the odds of hesitancy (OR 0.169, 95% CI 0.054-0.528; P=.002), a magnitude of effect not observed in higher SES groups. Additionally, a history of HZ infection was identified as a robust protective factor (OR 0.473, 95% CI 0.309-0.724; P=.001). Conclusions: eHealth literacy serves as a critical compensatory resource that can mitigate the disadvantage of low socioeconomic status in HZ vaccine acceptance. However, the protective mechanism is age-dependent, indicating a higher "digital threshold" for older adults (≥60 years) compared to their middle-aged counterparts. Public health interventions should prioritize digital empowerment for low-SES populations and tailor educational strategies to meet the higher digital competency needs of older adults. Clinical Trial: Not available

  • Background: Home spirometry has been widely adopted in the delivery of cystic fibrosis (CF) care. While existing literature largely supports its feasibility and positive outcomes, behaviour around home disease monitoring remains poorly understood. Objective: This study aimed to evaluate healthcare professionals’ (HCPs') ability to estimate home spirometry usage pwCF and compare these with actual recorded data. Methods: Home spirometry data, from a single large adult CF centre, for the year 2024, was obtained from NuvoAir. HCPs (doctors, nurses, and physiotherapists) rated their familiarity with each pwCF and categorised them as infrequent, expected, or highly frequent spirometry users. They were also asked to estimate spirometry usage as an open-ended numerical response. CF experience was defined by the number of years the HCP had worked at the centre. Estimation accuracy was assessed using mean bias and mean absolute error (MAE). Results: 10 doctors (35.7%), 6 nurses (21.4%), and 12 physiotherapists (42.9%) responded to the survey, with an overall response rate of 96.6%. There were 790 completed categorical estimates and 794 numerical estimates. The mean (±SD) CF experience was 15.7 (±8.2) years. Across all roles, HCPs systematically underestimated home spirometry usage (mean bias -4.9; MAE 6.32). No significant differences in estimation accuracy were observed based on professional role, reported familiarity or CF experience. Conclusions: This study found that CF caregivers tend to underestimate home spirometry usage, in contrast to other studies showing they often overestimate treatment adherence. This highlights gaps in understanding behaviour in pwCF and the need for CF teams to adapt to evolving models of remote monitoring.

  • Multidisciplinary Quality Control Circle–Based Workflow Optimization in Perioperative Respiratory Care: A Before-and-After Time–Motion Study

    Date Submitted: Feb 13, 2026
    Open Peer Review Period: Feb 14, 2026 - Apr 11, 2026

    Background: Perioperative respiratory care is a multidisciplinary process that includes several sequential steps and handoffs. Variability and inefficiencies within this workflow may delay care delivery and increase the workload of clinical staff. Quality Control Circles (QCCs) have been widely used in health care as a practical approach to addressing process-related quality problems identified in clinical practice. Objective: The aim of this study was to assess whether a Quality Control Circle–based intervention could improve workflow efficiency and care consistency in perioperative respiratory care. Methods: We conducted a before-and-after time–motion study on 30 perioperative respiratory care episodes to compare workflow before and after the implementation of a quality control circle. Recorded variables were: (1) total time from physician order to completion of respiratory care, (2) patient waiting time for incentive spirometry preparation, and (3) time clinical staff spent on patient education and respiratory training. Eligible patients were those prescribed perioperative respiratory care before or after surgery; those with prior exposure to the intervention or with hearing impairment were excluded. Guided by the Plan–Do–Check–Act cycle, improvement strategies included standardizing the provision of incentive spirometry, pre-positioning equipment at nursing stations, unifying education content, and delivering multimedia educational materials via quick response codes. Results: Before quality control circle implementation, the total process time was 266.65 minutes (240 minutes for equipment preparation and 17.5 minutes for patient education). After implementation, it dropped to 28.75 minutes (8.7 minutes for preparation, 10.5 minutes for schooling), improving overall efficiency by 89.2% and significantly reducing workflow time. Conclusions: A quality control circle framework not only optimized perioperative respiratory care but also engaged frontline staff, fostering a sense of teamwork and shared purpose. Multimedia patient education improved understanding and engagement, and cross-disciplinary collaboration reduced clinical workload. This strategy may reduce postoperative pulmonary complications and can be applied to other respiratory care workflows.

  • From Performance to Governance: An Adoption-Phase Ethical Perspective on Large Language Models in Health Care

    Date Submitted: Feb 13, 2026
    Open Peer Review Period: Feb 14, 2026 - Apr 11, 2026

    Large language models (LLMs) are increasingly integrated into everyday health-care communication, moving beyond experimental evaluation into routine clinical and informational use. Early research has primarily focused on technical performance, including accuracy, validation, and bias mitigation. While these remain essential, the transition to sustained real-world integration raises additional ethical questions that cannot be addressed by model-level evaluation alone. This Viewpoint proposes an adoption-phase ethics perspective, emphasizing how ethical risks shift as LLMs become embedded within institutional workflows, professional practices, and relationships of care. Drawing on normative analysis informed by existing empirical discussions, we examine three interrelated domains: trust, responsibility, and equity. During routine use, trust becomes shaped not only by perceptions of accuracy but also by expectations regarding accountability, transparency, and institutional protection. Responsibility may become diffused or ambiguous when LLM-mediated information influences clinical communication without clearly specified oversight. At the same time, differential digital literacy and access to institutional support may create uneven capacity to interpret and benefit from AI-generated information. We argue that ethical governance must therefore extend beyond pre-deployment technical safeguards toward sustained, system-level oversight. Adoption should be understood as a dynamic ethical process requiring role-sensitive design, clear accountability structures, and equity-oriented implementation. By reframing ethical attention from experimental validation to governance during routine integration, health-care systems can better ensure that the growing presence of LLMs supports fairness, responsibility, and patient trust alongside technical advancement.

  • Background: Fitness and physical activity patterns are key predictors of cardiovascular disease. Traditionally, these factors have been assessed through participant self-report, which is prone to recall bias and inaccuracy. Smartphone-based monitoring provides a scalable and objective alternative for measuring physical activity, offering improved accuracy over conventional assessment methods. Objective: To evaluated the feasibility of smartphone-based cardiovascular research in the Netherlands and to examine associations between objectively measured physical activity, perceived activity, functional capacity, life satisfaction, and cardiovascular risk. Methods: Adults in the Netherlands were recruited via the MyHeart Counts iPhone app between August 2022 and December 2023. Within the app, participants completed surveys, passively shared motion sensor data, and were invited to perform a smartphone-based 6-minute walk test (6MWT). Perceived activity was compared with sensor-measured activity and actual activity (sensor-measured with supplemented self-reported unrecorded activity). Multivariable linear regression assessed associations between activity and 6MWT performance and between activity and life satisfaction. Perceived cardiovascular risk was compared with the difference between heart age and actual age. Results: Of 518 enrolled participants (median age 58 years; 72% female), 93% shared data beyond demographics. Median engagement duration was 27 days, and 58% completed at least one full consecutive week of motion tracking. Perceived activity weakly correlated with both sensor-measured activity (ρ = 0.15, P = .01) and with actual activity (ρ = 0.15, P = .01). Median perceived activity was 3.5 hours/week, significantly higher than sensor-measured activity (0.9 hours/week; mean difference 2.9 hours, 95% CI 2.2–3.7; P < .001). In contrast, median actual activity was 3.2 hours/week and did not differ significantly from perceived activity (mean difference 0.7 hours, 95% CI −0.2 to 1.6; P = .11), indicating no significant over- or underestimation when unrecorded activity was accounted for. Sensor-measured physical activity was associated with longer 6MWT distance (+10.1 m per hour; 95% CI 3.9-16.4, P = 0.002). No association was observed between sensor-measured activity and life satisfaction. Perceived cardiovascular risk correlated with the difference between heart age and actual age (ρ = 0.41; P < 0.001). Conclusions: Smartphone-based cardiovascular monitoring is feasible in a European adult population and yields valid functional correlates of physical activity. However, incomplete phone carriage substantially limits sensor-only activity estimates, underscoring the need for hybrid measurement strategies. These findings support the use of smartphone platforms for scalable cardiovascular research, while highlighting persistent challenges in engagement and measurement completeness.

  • Background: While AI's transformative potential in healthcare is widely acknowledged, its application in highly sensitive, humanistic domains like PPC remains largely unexplored. Objective: To explore the attitudes and needs of healthcare providers on the pediatric palliative care (PPC) assisted by artificial intelligence (AI), with the goal of informing future development and implementation of AI systems in this field. Methods: This was an explanatory sequential mixed-methods study consisting of a nationwide cross-sectional questionnaire survey (March–April 2025) followed by qualitative semi-structured interviews (August–October 2025). The quantitative study aimed to investigate PPC healthcare providers' experiences, attitudes, and needs for the application of AI. Participants included team members of all recognized PPC teams in mainland China. The qualitative study aimed to explore in greater depth the potential future roles of AI in this field, as well as the features of an ideal AI-assisted tool for PPC. Potential interviewees were recruited from the pool of quantitative survey respondents. Results: Among 352 survey respondents, most (58.24%) reported moderate familiarity with AI, with large language models being the most commonly used (79.55%). Among large language model users, over half (57.50%) reported using them for clinical purposes. Attitudes were generally positive: 67.05% believed AI's benefits would outweigh drawbacks, and 78.98% considered its implementation feasible. The most desired applications were patient/family education (78.41%) and symptom management (73.01%). Interviews with 17 providers revealed three themes: (1) clinical roles and boundaries; (2) elements for clinical integration; and (3) challenges in development and deployment. Conclusions: This study reveals that PPC providers express positive attitudes and strong demand for AI-assisted clinical work. Furthermore, the research clarifies appropriate roles for AI, outlines elements for clinical integration, and highlights potential challenges in development and integration. This study provides evidence for the feasibility of AI application in PPC and offers guidance for the future development and deployment of AI tools.

  • Semantic Layer in Health Care: The Art of Riding a Bicycle

    Date Submitted: Feb 10, 2026
    Open Peer Review Period: Feb 11, 2026 - Apr 8, 2026

    Health data interoperability is the central hill climb in contemporary digital health. Hospitals often accumulate data like mismatched spare parts, catalogued inconsistently, and difficult to re-use across care. The landscape of non-annotated source systems, legacy data warehouses that lack interoperable data models, the coexistence of multiple terminologies with divergent scopes, the operational turbulence of system migrations, and the persistent challenges of metadata catalogues and versioning set a starting point to a journey in building a semantic layer that makes data Findable, Accessible, Interoperable, and Reusable (FAIR), and that remains robust as terminologies evolve. Terminology updates are complex and their terms, classifications, and regulations continually change. This viewpoint article gives an exemplary historical overview at a Swiss university hospital, highlights the relevance of key decisions and projects and contrasts local conditions with the Swiss and European context. It notes perspectives of large clinical information systems and highlights organizational implications, tools and models needed, and the challenge of legacy data. It dives into project work of ontology creation. The discussion reflects on achievements and the future illustrating the cadence and resilience required to ride interoperable data “around the world”. Key Message. Achieving healthcare interoperability requires balancing diverse standards, terminologies, and data governance. The FAIR principles provide a framework. Organizational commitment to these practices is essential.

  • Public Online Discussions of CAR T-cell Cancer Therapy: Unpacking the Hype

    Date Submitted: Feb 9, 2026
    Open Peer Review Period: Feb 10, 2026 - Apr 7, 2026

    Background: Chimeric antigen receptor (CAR) therapy is a novel cell editing technology and innovative form of cancer immunotherapy. An individual’s immune cells (T-cells) are removed from the body, engineered to target and limit the growth of cancer cells, and reinfused into the patient’s body. The one-time treatment is expensive ($500,000 plus hospital costs), and requires specialized care to treat and manage the associated side effects, such as cytokine release syndrome (CRS), and other serious health issues including cognitive confusion, infertility, secondary malignancies, and compromised long term quality of life. At the same time, CAR T has been highly successful for patients with advanced blood cancers and no remaining treatment options. The CAR T landscape is changing rapidly, and product approvals have outpaced the capacity for researchers to collect long term evidence related to survival or predictive biomarkers that might better prioritize patients. Because CAR T is offered exclusively in urban cancer centres with access to cell manufacturing capacity, equitable access has been challenging. At the same time there is considerable demand and social hype about CAR T as a cancer cure despite the risks and uncertainty of the technology. Objective: We aimed to determine the dominant perspectives and nature of the information on CAR T-cell therapy available to the public in the online environment. Methods: In this qualitative study, we conducted a comprehensive search of websites including professional, medical, corporate, health-based, news media, and blogs to capture the diversity of online sources and their perspectives presenting information on CAR T-cell therapy. Fifty-one webpages met the study criteria and comprised the data set in this review. The content of the sites was reviewed and analyzed using a critical and interpretive descriptive lens. Results: We classified the website information into four dominant major themes characterizing CAR T-cell therapy: 1) patient stories of success, magic and hope; 2) medical science explainers; 3) economic perspectives; and 4) ethical discussions and complex arguments. With the exception of the sites that presented ethical discussions and complex information, the online environment positioned CAR T as revolutionary, curative, and the future of cancer treatment. Side effects were generally minimized, and collective dilemmas such as sustainability for the healthcare system, equitable access, and issues of prioritization were frequently sidelined or absent. Conclusions: The persuasive tone of online CAR T information combined with the increasingly blurred distinctions between research and care in genetic medical technologies suggests that obtaining informed consent or refusal may place too much onus on individual patients. In an evolving technological landscape such as CAR T, determining the acceptable risks and benefits is a question that ethically requires broader, as well as more inclusive, societal deliberation.