Explore the vast range of titles available.

The global public health challenge of young people's mental health is particularly evident in Brazil, which is considered the most anxious and the fifth most depressed country globally. In response to this, a digital intervention, that takes the form of a storytelling chatbot named \"Cadê o Kauê?\" (translation: \"Where is Kauê?\"), was coproduced by an interdisciplinary team of researchers, young people, and narrative designers as part of a project titled Engajadamente. The intervention was designed to strengthen Brazilian adolescents' skills in promoting their peers' mental health through direct support and collective action for mental health. The intervention was rolled out in select schools in Brazil and online via social media, but it encountered challenges during the initial implementation phase in school settings. This project's primary objective is to evaluate the implementation of \"Cadê o Kauê\" in school settings in Brazil, assess current and potential complexities, and formulate recommendations to reduce or manage these complexities. The secondary objective of this project is to understand the current and potential sociotechnical change resulting from the implementation of the \"Cadê o Kauê\" in a real-world setting. The Non-adoption, Abandonment, Scale-up, Spread, Sustainability (NASSS) framework was selected as the theoretical framework to identify different areas of complexity across domains and further thinking of ways to reduce or manage these complexities. Data were gathered by conducting semistructured online interviews with the Engajadamente team members based in the United Kingdom and Brazil. The NASSS-Complexity Assessment Tool was used as a sensemaking device to facilitate a collective understanding of the implementation, and narratives were mapped onto specific NASSS domains. Key themes were identified across overarching domains using thematic analysis to meet the objectives of this evaluation. The results mapped onto the NASSS domains generated a narrative about the initial implementation of \"Cadê o Kauê?\" in schools. Existing and potential complexities across technology, organization, and wider system domains were characterized by interdependencies, unintended consequences, and uncertainties. The implementation of \"Cadê o Kauê?\" highlighted challenges related to internet dependency, infrastructure limitations, and varying levels of organizational readiness, teacher preparedness, and school culture. Addressing these through offline solutions, additional teacher training, whole-school engagement strategies, and a value-informed complexity approach can enhance accessibility, alignment with school priorities, and adaptability within Brazil's dynamic sociopolitical context. Using the NASSS framework, the evaluation captured both current and emerging complexities in implementing \"Cadê o Kauê?\" in schools. By engaging with theory-informed approaches to technology implementation projects and deep-diving into discursive dynamic interactions between technology and its \"context\" of implementation, we have explored ways to manage these complexities and developed suitable recommendations to guide the Engadajamente team and support similar projects in the future.

Primary health care (PHC) is critical for delivering accessible and continuous care but faces persistent challenges such as workforce shortages, administrative burden, and rising multimorbidity. Artificial intelligence (AI) has the potential to support PHC by enhancing diagnosis, workflow efficiency, and clinical decision-making. However, existing research often overlooks how AI tools function within the complex realities of primary care and how clinicians and patients experience them. This scoping review maps the landscape of AI applications in PHC, with a focus on empirical studies involving direct engagement from PHC stakeholders. The review emphasizes real-world settings, clinical workflows, and the alignment of AI tools with the values and complexity of generalist care. Following Joanna Briggs Institute methodology and PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) guidelines, we searched PubMed, Web of Science, and Scopus databases up to April 13, 2024. Inclusion criteria were empirical, peer-reviewed studies published in English between January 2010 and April 2024, involving direct stakeholder interaction (general practitioners, nurses, or patients) in real-world PHC settings, evaluating AI applications (eg, diagnostics, workflow optimization, and documentation). Exclusions included algorithm-only validations, pediatric populations, secondary or tertiary care contexts not explicitly addressing PHC workflows, nonempirical research (eg, editorials or protocols), and non-English studies. We used thematic analysis to synthesize findings related to study aims, AI applications, and stakeholder roles. Of 5224 identified records, 73 studies met the inclusion criteria. Studies were grouped into four main themes: (1) early intervention and decision support (n=21; 29%), (2) chronic disease management (n=16; 22%), (3) operations and patient management (n=12; 16%), and (4) acceptance and implementation experiences (n=24; 33%). AI tools frequently demonstrated strong technical accuracy, particularly in diagnostic decision support. However, implementation in routine practice was often limited by usability barriers, workflow misalignment, trust concerns, equity gaps, and financial constraints. Overall, AI holds significant potential to support PHC, especially when aligned with clinical reasoning, workflow needs, and relational care models. However, persistent implementation barriers such as usability challenges, training gaps, and workflow integration issues must be addressed. The evidence included in this review is limited by heterogeneity in study design and the predominance of small-scale feasibility studies. Future research should prioritize pragmatic trials, co-design with PHC professionals, and anticipatory planning using future methods to ensure responsible and equitable implementation.

The fast growth of social media mining in health research has contributed to an invaluable but quite fragmented body of literature. As the amount of unstructured patient-reported data grows, traditional bibliometric analyses face methodological limitations, particularly regarding synonym fragmentation and arbitrary parameter selection. In their recent publication, \"Thematic Mapping and Evolution of Social Media Mining in Health Research: Hybrid Bibliometric Synthesis,\" Yang and Bohnet-Joschko attempt to address these flaws by introducing a semantic-structural (hybrid) bibliometric framework. This commentary evaluates the methodological innovations of their study and its departure from traditional syntactic keyword-matching tools. By combining citation-informed transformers (SPECTER2) and biomedical language models (PubMedBERT) and dimensionality reduction and density-based clustering, the authors created a reproducible pipeline. In their architecture, they start with foundational machine learning (statistical validity) before transitioning into large language models for qualitative synthesis. I will attempt to explain how this transition from syntactic mapping to semantic vector representation solves known challenges in evidence synthesis, naturally grouping conceptual synonyms without artificially forcing boundaries on the literature. Furthermore, I examine the practical implications of their temporal findings. Such real-time social media mining applications can be very useful for retrospective reporting and evaluating targeted public health interventions. While this pipeline offers high generalizability across disciplines, it also introduces a computational literacy barrier to some, and this re-emphasizes the need for data literacy for health professions. Ultimately, the study provides a transparent approach to informatics because mathematically validated frameworks are foundational for the future of evidence-driven public health policy and clinical decision-making.

Rapidly and accurately synthesizing large volumes of evidence is a time- and resource-intensive process. Once published, reviews often risk becoming outdated, limiting their usefulness for decision makers. Recent advancements in artificial intelligence (AI) have enabled researchers to automate stages of the evidence synthesis process, from literature searching and screening to data extraction and analysis. As previous reviews on this topic have been published, a significant number of tools have been further developed and evaluated. Furthermore, as generative AI increasingly automates evidence synthesis, understanding how it is studied and applied is crucial, given both its benefits and risks. This review aimed to map the current landscape of evaluated AI tools used to automate evidence synthesis. Following the Joanna Briggs Institute methodology for scoping reviews, we searched Ovid MEDLINE, Ovid Embase, Scopus, and Web of Science in February 2025 and conducted a gray literature search in April 2025. We included articles published in any language from January 2021 onward. Two reviewers independently screened citations using Rayyan, and data were extracted based on study design and key AI-related technical features. We identified 7841 unique citations through database searches and 19 records through gray literature searching. A total of 222 articles were included in the review. We identified 65 AI tools and 25 open-source models or machine learning (ML) algorithms that automate parts of or the whole evidence synthesis pathway. A total of 54.1% (n=120) of the studies were published in 2024, reflecting a trend toward researching general-purpose large language models (LLMs) for evidence synthesis automation. The most popular tool studied was generative pretrained transformer models, including its conversational interface ChatGPT (n=70, 31.5%). Moreover, 31.1% (n=69) studied tools automated by traditional ML algorithms. No studies compared traditional ML tools to LLM-based tools. In addition, 61.7% (n=137) and 26.1% (n=58) studied AI-assisted automation of title and abstract screening and data extraction, respectively, the 2 most intensive stages and, therefore, amenable to automation. Technical performance outcomes were the most frequently reported, with only 4.1% (n=9) of studies reporting time- or workload-specific outcomes. Few studies pragmatically evaluated AI tools in real-world evidence synthesis settings. This review comprehensively captures the broad, evolving suite of AI automation tools available to support evidence synthesis, leveraged by increasingly complex AI approaches that range from traditional ML to LLMs. The notable shift toward studying general-purpose generative AI tools reflects how these technologies are actively transforming evidence synthesis practice. The lack of studies in our review comparing different AI approaches for specific automation stages or evaluating their effectiveness pragmatically represents a significant research gap. Optimal tool selection will likely depend on the review topic and methodology and researcher priorities. While they offer potential for reducing workload, ongoing evaluation to mitigate AI bias and to ensure the integrity of reviews is essential for safeguarding evidence-based decision-making.

Evaluating implementation of digital health interventions (DHIs) in practice settings is complex, involving diverse users and multistep processes. Proactive planning can ensure implementation determinants and outcomes are captured for hybrid studies, but operational guidance for designing or planning hybrid DHI studies is limited. This study aimed to proactively define, prioritize, and operationalize measurement of implementation outcomes and determinants for a DHI hybrid effectiveness-implementation trial. We describe unique advantages and limitations of planning the trial implementation evaluation among a large-scale cohort study population and share results of a pretrial organizational readiness assessment. We planned a cluster-randomized, type II hybrid effectiveness-implementation trial testing PositiveLinks, a smartphone app for HIV care, compared to usual care (n=6 sites per arm), among HIV outpatient sites in the DC Cohort Longitudinal HIV Study in Washington, DC. We (1) defined components of the DHI and associated implementation strategy; (2) selected implementation science frameworks to accomplish evaluation aims; (3) mapped framework dimensions, domains, and constructs to implementation strategy steps; (4) modified or created instruments to collect data for implementation outcome measures and determinants; and (5) developed a compatible implementation science data collection and management plan. Provider baseline surveys administered at intervention sites probed usage of digital tools and assessed provider readiness for implementation with the Organizational Readiness to Implement Change tool. We specified DHI and implementation strategy toward planning measurement of DHI and broader program reach and adoption. Mapping of implementation strategy steps to the Reach Effectiveness Adoption Implementation Maintenance framework prompted considerations for how to capture understudied aspects of each dimension: denominators and demographic representativeness within reach or adoption, and provider or organization-level adaptations, dose, and fidelity within the implementation dimension. Our process also prompted the creation of tools to obtain detailed determinants across domains and constructs of the Consolidated Framework for Implementation Research within a large sample at multiple time points. Some aspects of real-world PositiveLinks implementation were not reflected within the planned hybrid trial (eg, research assistants selected as de facto site implementation leads) or were modified to preserve internal validity of effectiveness measurement (eg, \"Community of Practice\"). Providers and research assistants (n=17) at intervention sites self-reported high baseline use of digital tools to communicate with patients. Readiness assessment revealed high median (48, IQR 45-54) total Organizational Readiness to Implement Change scores, with research assistants scoring higher than physicians (52.5, IQR 44-55 vs 48.0, IQR 46-49). Key takeaways, challenges, and opportunities arose in planning the implementation evaluation within a hybrid DHI trial among a cohort population. Prospective trial planning must balance generalizability of implementation processes to \"real world\" conditions with rigorous procedures to measure intervention effectiveness. Rapid, scalable tools require further study to enable evaluations within large multisite hybrid studies.

Diagnosing rare diseases remains challenging due to their inherent complexity and limited physician knowledge. Large language models (LLMs) offer new potential to enhance diagnostic workflows. This study aimed to evaluate the diagnostic accuracy of ChatGPT-4o and 4 open-source LLMs (qwen2.5:7b, Llama3.1:8b, qwen2.5:72b, and Llama3.1:70b) for rare diseases, assesses the language effect on diagnostic performance, and explore retrieval augmented generation (RAG) and chain-of-thought (CoT) reasoning. We extracted clinical manifestations of 121 rare diseases from China's inaugural rare disease catalog. ChatGPT-4o generated a primary and 5 differential diagnoses, while 4 LLMs were assessed in both English and Chinese contexts. The lowest-performing model underwent RAG and CoT re-evaluation. Diagnostic accuracy was compared via the McNemar test. A survey evaluated 11 clinicians' familiarity with rare diseases. ChatGPT-4o demonstrated the highest diagnostic accuracy with 90.1%. Language effects varied across models: qwen2.5:7b showed comparable performance in Chinese (51.2%) and English (47.9%; χ²1=0.32, P=.57), whereas Llama3.1:8b exhibited significantly higher English accuracy (67.8% vs 31.4%; χ²1=40.20, P<.001). Among larger models, qwen2.5:72b maintained cross-lingual consistency considering the odds ratio (OR; Chinese: 82.6% vs English: 83.5%; OR 0.88, 95% CI 0.27-2.76,P=1.000), contrasting with Llama3.1:70b's language-dependent variation (Chinese: 80.2% vs English: 90.1%; OR 0.29,95% CI 0.08-0.83, P=.02). Cross-model comparisons revealed Llama3.1:8b underperformed qwen2.5:7b in Chinese (χ²1=13.22,P<.001) but surpassed it in English (χ²1=13.92,P<.001). No significant differences were observed between qwen2.5:72b and Llama3.1:70b (English: OR 0.33, P=.08; Chinese: OR 1.5, 95% CI 0.48-5.12,P=.07); qwen2.5:72b matched ChatGPT-4o's performance in both languages (English: OR 0.33, P=.08; Chinese: OR 0.44, P=.09); Llama3.1:70b mirrored ChatGPT-4o's English accuracy (OR 1, P=1.000) but lagged in Chinese (OR 0.33; P=.02). RAG implementation enhanced qwen2.5:7b's accuracy to 79.3% (χ²1=31.11, P<.001) with 85.9% retrieval precision. The distilled model Deepseek-R1:7b markedly underperformed (9.9% vs qwen2.5:7b; χ²1=42.19, P<.001). Clinician surveys revealed significant knowledge gaps in rare disease management. ChatGPT-4o demonstrated superior diagnostic performance for rare diseases. While Llama3.1:8b demonstrates viability for localized deployment in resource-constrained English diagnostic workflows, Chinese applications require larger models to achieve comparable diagnostic accuracy. This urgency is heightened by the release of open-source models like DeepSeek-R1, which may see rapid adoption without thorough validation. Successful clinical implementation of LLMs requires 3 core elements: model parameterization, user language, and pretraining data. The integration of RAG significantly enhanced open-source LLM accuracy for rare disease diagnosis, although caution remains warranted for low-parameter reasoning models showing substantial performance limitations. We recommend hospital IT departments and policymakers prioritize language relevance in model selection and consider integrating RAG with curated knowledge bases to enhance diagnostic utility in constrained settings, while exercising caution with low-parameter models.

Close follow-up of stable patients with axial spondyloarthritis (axSpA) presents a financial burden and inconvenience to patients. A remote monitoring patient-reported outcome measures (PROMs)-based model of care (PROMise) was designed to reduce the frequency of in-person consultations for stable patients with axSpA. However, little is known about the facilitators and barriers of implementing a remote monitoring PROMise. This study aims to understand the facilitators and barriers, as well as the mitigation strategies to implementing a PROMise in the Singapore context. We conducted a qualitative study involving in-depth interviews with 19 patients with axSpA (78.9% (15) male, mean age 39.4, SD 11.7 years) and 13 health care professionals (HCPs) (23.1%, 3 male; mean age 37.9, SD 7.2 years) in a tertiary hospital in Singapore until data saturation was reached. Participants were purposively recruited based on sex, age, and ethnicity. Patients were additionally recruited based on the number of years since diagnosed with axSpA, while HCPs were recruited based on seniority and their role in the care of patients with axSpA. Interviews were transcribed, deductively analyzed, and mapped to the Consolidated Framework for Implementation Research (CFIR) framework to identify facilitators and barriers from both the patients' and HCPs' perspectives. The CFIR-Expert Recommendations for Implementing Change (ERIC) match tool was used to produce implementation strategies to overcome the CFIR barriers identified. All five domains of the CFIR framework were elicited. Facilitators included (1) reduced inconvenience and costs for patients and reduced patient load in the clinic, (2) need for PROMise, (3) similarity to current workflows, and (4) suitable patient selection. Barriers included concerns for (1) financial sustainability of PROMise, (2) cultural conditions, (3) patient safety, and (4) increased workload for HCPs. In total, 35 ERIC strategies were matched to the corresponding CFIR barriers. We identified ERIC strategies that will facilitate the implementation of the PROMise model. In particular, focus should be placed on developing an implementation blueprint and obtaining continuous feedback from affected patients with axSpA and HCPs involved in the care of the affected patients. These implementation strategies cross-cut the CFIR barriers identified and thus may overcome the barriers to implementation.

Health care access plays a central role in reducing inequities across populations. Virtual care can mitigate these inequities by facilitating more inclusive and accessible health care delivery. In residential aged care homes (RACHs), virtual care has the potential to enable timely and efficient access to general practitioners (GPs) for residents. However, as context, technologies, and users are complex, the implementation, acceptance, and use of virtual care technologies in RACHs remain challenging. This study aimed to evaluate the barriers and facilitators to implementing, accepting, and using virtual care technologies to connect residents with GPs for health care delivery in RACHs, and to identify benefits, unintended consequences, and opportunities for optimization. We conducted a pragmatic, cross-sectional qualitative study guided by interpretivist principles. Semistructured interviews were undertaken with 32 participants (11 GPs, 11 RACH nurses, 3 practice managers, 5 residents, and 2 carers). Data were analyzed inductively using reflexive thematic analysis and mapped deductively to the Systems Engineering Initiative for Patient Safety model to examine sociotechnical interactions influencing virtual care delivery. Our investigation revealed that barriers to implementing, accepting, and using virtual care technologies to deliver care to RACH residents were more pervasive and salient in participants' accounts than enablers. While barriers were found across all Systems Engineering Initiative for Patient Safety domains for GPs, most of the barriers for residents and carers were identified in the \"people\" and \"organizational\" domains, and in addition to these, \"technology and tools\" domain for RACH nurses. Although many barriers are common across the people (eg, resistance to using new technology), technology (eg, inadequate system integration), and organization (eg, logistical challenges) domains for RACH nurses and GPs, our study revealed unique barriers to virtual care delivery for residents and carers (eg, interruptions and potential to exclude residents from conversations) whose views are often absent from existing literature. Our findings also revealed that there is no standardized virtual care consultation process between RACHs and GPs-a key concern strongly associated with the identified work system barriers. While virtual care was seen as beneficial, participants identified some unintended consequences to patients (eg, loss of doctor-patient relationship), clinicians (eg, additional workload), and health care organizations (eg, infection control). Virtual care can improve access to timely, high-quality general practice services in RACHs, but its potential is constrained by sociotechnical, organizational, and workflow challenges. Addressing system integration, usability, staffing, training, and policy gaps, including funding and billing structures, will be crucial to realizing the benefits of virtual care. This study provides new evidence to inform design, implementation, and policy decisions supporting equitable virtual care delivery aligned with Sustainable Development Goals 3 and 10.

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. Health care 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. This study aims 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. 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 GenAI 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. 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 themes. 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. 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.

Applying digital health technologies (DHTs) for health promotion and disease prevention is recommended by official bodies such as the World Health Organization. User-centered co-design with systematic patient and public involvement is considered best practice for developing such complex interventions. Although well-established methodological guides and frameworks are available, an important gap is that they are either holistic but generic, offering minimal operational guidance, or context-specific and operational, but focusing only on subphases of establishing DHT-enhanced interventions. This paper presents a unifying consensus-based methodological guideline directed toward multidisciplinary expert teams coordinating projects on individually tailored DHTs. It delineates best practices with operational guidance for each step along the full lifecycle of DHT-enhanced training and rehabilitation concepts-from contextualization, through codevelopment, and evaluation to implementation. The Co-Develop-IT guideline was cocreated through a structured expert consensus process that integrated, refined, and expanded on well-established existing guides and frameworks to delineate holistic and context-specific, yet flexible enough, best practices. The process consisted of biweekly 90-minute hybrid meetings between August 2024 and February 2025, in combination with written elaboration, feedback, and revisions between meetings to gradually develop a consensus on best practice recommendations. The Co-Develop-IT guideline consists of 8 iterative phases. It is applicable to any type of end users, exercise types, intended contexts of use (eg, primary health care, community health services, and telemedicine), and overarching goals (eg, health promotion and primary through tertiary disease prevention, including rehabilitation). The Co-Develop-IT guideline introduces 5 distinct preparatory contextual research phases preceding generative codevelopment. These phases are dedicated to the structured establishment of a more robust foundation to better tailor and steer codevelopment efforts toward successful implementation. In 2 application examples, we provide proof of concept that the resulting guideline fulfills its primary purpose of providing comprehensive, context-specific, and operational, yet flexible enough best practice recommendations. The unifying Co-Develop-IT guideline provides comprehensive best practices with actionable operational guidance for establishing an appropriate balance between scientific theories and frameworks and the real-world needs of interest-holders in the establishment of individually tailored DHT-enhanced training and rehabilitation concepts. Applying Co-Develop-IT contributes to overcoming the lingering evidence-to-practice gap by consistently establishing a shared mission with relevant interest-holders and ensuring that all codevelopment steps are directed toward addressing an unmet need in (clinical) practice-ultimately promoting the practical application and impact of purpose-developed DHTs.