Healthcare

IntroductionLarge language models (LLMs) have gained traction in medicine, but there is limited research comparing closed- and open-source models in subspecialty contexts. This study evaluated ChatGPT-4.0o and DeepSeek-R1 on a pediatric cardiology board-style examination to quantify their accuracy and discuss clinical and educational utility. MethodsChatGPT-4.0o and DeepSeek-R1 were used to answer 88 text-based multiple-choice questions across 11 pediatric cardiology subtopics from a Pediatric Cardiology Board Review textbook. DeepSeek-R1s processing time per question was measured. Statistical analyses for model comparison were conducted using an unpaired two-tailed t-test, and bivariate correlations were assessed using Pearsons r. ResultsChatGPT-4.0o and DeepSeek-R1 achieved 70% (62/88) and 68% (60/88) accuracy, respectively (p=0.79). Subtopic accuracy was equal in 5 of 11 chapters, with each model outperforming its counterpart in 3 of 11. DeepSeek-R1s processing time negatively correlated with accuracy (r = -0.68, p = 0.02). ConclusionChatGPT-4.0o and DeepSeek-R1 approached the passing threshold on a pediatric cardiology board examination, with comparable accuracy and potential for open-source models to enhance clinical and educational outcomes while supporting sustainable AI development.

BackgroundAlzheimers disease and related dementia (ADRD) in the United States (U.S.) disproportionately affects African Americans (AA). Previous research has indicated that healthy eating habits help slow cognitive decline among adults 65 years and older. However, dietary intervention studies that demonstrated preliminary ability to protect against cognitive decline demonstrated low adherence and acceptability among African Americans. ObjectiveThe objective of this study is to identify and understand knowledge and attitudes that influence dietary practices among older African Americans using a community-engaged approach. DesignA non-interventional mixed methods study designed to inform the development of an adapted brain-healthy soul food diet intervention. A purposive sampling approach was used to conduct 5 semi-structured focus group discussions and an online quantitative survey. The Health Belief Model guided the focus group using the following constructs: susceptibility, perceived severity, perceived benefits, perceived barriers, self-efficacy, and cue to action. Participants/settingInclusion criteria included self-identifying as African American or Black, aged 55 years and older, English proficient, and cognitively normal with an AD8 < 2. Participant were asked to participate in a single 60-minute virtual focus group discussion. In total, 39 individuals (25.6% men, 74.4% women) took part in one of the seven virtual focus group discussion (5-7 per focus group). Main outcome measureKnowledge and attitudes that influence dietary practices among older African Americans. Statistical analysis performedFocus groups were analyzed using a 6-step thematic analysis process, and quantitative survey data was analyzed using descriptive statistics. ResultsFive themes emerged: Knowledge of dementia; practices shaping food choices and consumption; barriers impacting healthy dietary consumption; experiential instrumental support; and elements of a culturally tailored brain healthy dietary intervention. ConclusionOlder Midwesterners perceived an adapted MIND dietary model as most feasible with the incorporation of salient cultural characteristics and strategies within both the design and delivery phases. Research SnapshotO_ST_ABSResearch QuestionC_ST_ABSWhat are the perceptions related to healthy diets and the facilitators and barriers that influence dietary practices among older Midwestern African Americans? Key FindingFive themes emerged from the focus group discussions: Knowledge of dementia; Practices shaping food choices and consumption; Barriers impacting healthy dietary consumption; External instrumental support; and Elements of culturally tailored brain healthy dietary intervention. Facilitators identified to support following a healthy diet included cooking education, food preparation demonstrations, and accessibility guidance. Barriers identified that impact healthy dietary consumption included access, cost, taste, and food spoilage.

Preterm birth is a common neurodevelopmental condition that can have lasting impacts on cognition, including attention and working memory. Interventions that strengthen these skills in early childhood could support school readiness. Dino Island is a tablet-based intervention that combines process-specific practice of attention and working memory skills with a compensatory component that teaches metacognitive skills to scaffold learning. In this pilot study, we evaluated the feasibility of parent-delivered Dino Island in young children born preterm (N=10), alongside a control group who played educational games (N=12). Both groups were instructed to play 2-3 times per week for approximately 20 minutes over a 12-week period. Attention and working memory on untrained tasks were assessed before and after completing the intervention. Parents provided fidelity data through tracking sheets and participated in exit interviews to offer feedback and identify barriers and facilitators. We found that the Dino Island program was successfully delivered by parents with high fidelity. Attrition was higher in the Dino Island group, likely reflecting the challenges of delivering cognitive remediation in home settings. Comparison of attention and working memory scores on untrained tasks pre- and post-showed practice effects but no specific benefit of Dino Island. However, parent reports suggested behavioral improvements specific to the Dino Island group, noting far-transfer effects where children applied metacognitive strategies in other contexts. Overall, this work shows feasibility and tolerability of Dino Island in young children born preterm. Future research should examine its potential impact on school readiness and longer-term academic outcomes in this population.

BackgroundCardiac amyloidosis (CA) is a rare, progressive condition that requires complex management, highlighting the importance of accessible and understandable patient education materials (PEMs). The American Medical Association (AMA) recommends that PEMs be written at or below a 6th-grade reading level; however, PEMs for CA often exceed this recommendation. GPT-4, a large language model (LLM), is increasingly being studied for its ability to enhance PEM readability. Materials and MethodsInstitutional PEMs were sourced from the websites of the top 10 institutions ranked by the 2022-2023 US News & World Report for "Best Hospitals for Cardiology, Heart & Vascular Surgery." GPT-4 (version updated 20 July 2023) was prompted with "please explain the following in simpler terms" along with each institutional PEM to produce revised responses. The readability of both the institutional and GPT-4 revised PEMs was evaluated using validated readability formulas through readable.com. Finally, a board-certified cardiologist reviewed institutional and revised PEMs to assess for changes in accuracy and comprehensiveness. ResultsA total of 86 PEMs were analyzed. None of the institutional PEMs met the recommended 6th-grade reading level and had a median Flesch-Kincaid Grade Level (FKGL) of 10.9 (high school freshman; IQR: 9.2, 12.6, p<0.001). In comparison, GPT-4 revised PEMs reduced the median FKGL to 7.8 (seventh grade; IQR: 7.0, 8.8, p<0.001), with 14/86 (16.3%) achieving at least a 6th-grade reading level. All GPT-4 revised PEMs 86/86 (100%) were accurate after review, with 9/86 (10.5%) deemed more comprehensive and 3/86 (3.5%) deemed less comprehensive than the institutional materials. ConclusionsGPT-4 significantly improved the readability of institutional PEMs for CA while maintaining accuracy and, in some cases, enhancing comprehensiveness. These findings underscore the potential of LLMs to bridge health literacy gaps by simplifying complex medical information without compromising content integrity. However, further research is needed to validate our findings and assess patient comprehension, real-world efficacy, and the impact of AI-driven education on clinical outcomes.

Breastfeeding has many benefits for the newborn and mother therefore, the World Health Organization guidelines recommend exclusive breastfeeding for the first 6 months, with continued breastfeeding for up to two years. However, exclusive breastfeeding rates in the first 6 months in Thailand were reported to be approximately, 14% in 2019. Research has highlighted that many mothers have a lack of belief in their ability to breastfeed, and some physical conditions, such as tiredness and difficulty continually holding their baby in a comfortable position. Additionally, first-time breastfeeding can contribute to mothers having difficulties breastfeeding during the early postnatal period. Therefore, the Arm sling innovation (device) has been designed to provide support and comfort whilst holding the newborn. This study compared the effectiveness of the breastfeeding arm sling innovation to support breastfeeding in cross-cradle hold position and normal cross-cradle hold position breastfeeding in first-time mothers. A quasi-experimental pretest-posttest research design was used to evaluate the effectiveness of breastfeeding before and after the intervention among first-time mothers in the postnatal unit, at Ramathibodi hospital, Thailand. A total of 46 postpartum mothers participated in the study. The results showed that the effectiveness of breastfeeding reported by mothers between using a normal cross-cradle hold position and using breastfeeding arm sling innovation was statistically significantly different (t = 4.32, P < 0.001) with helping to hold the baby securely without slipping (t=5.68, p<0.001) and mothers can continue to breastfeed (t=2.09, p <0.001). Majority of mothers were satisfied with the support of using the breastfeeding arm sling innovation design. The breastfeeding arm sling innovation contributes to the effectiveness of breastfeeding by assisting and supporting the mother and babys position to breastfeed more comfortably, thus assisting first-time mothers to feel comfortable, confident, and able to continue breastfeeding.

BackgroundThe quality of healthcare services has garnered significant attention within the medical field, with patient-perceived satisfaction emerging as a crucial indicator. Analyzing patient satisfaction could uncover areas of weakness, suggesting the need for improvements in care. However, in the specialized field of Fetal Medicine and Therapy, there lacks a validated tool for assessing care quality from the patients perspective. This study aims to design and pilot an instrument for measuring patient-perceived satisfaction as a barometer of the quality of services provided. MethodsWe developed a survey instrument to assess six key areas of healthcare (Follow-up care, environment, transparency, accessibility of care, link between physician and patient, and global satisfaction), through 20 multiple-choice questions. It was initially piloted on a sample of 34 pregnant women undergoing a specific fetal therapy (Ex Utero Intrapartum Treatment, EXIT) at our center between June 2007 and January 2024. ResultsA total of 29 out of 34 patients agreed to participate. According to the proposed tool and the suggested scoring scale, we achieved an overall score of 4.67, indicating a very satisfactory rating. The highest mean score was for the "Global satisfaction" domain (4.76), revealing that participants where highly likely to recommend our department. It was followed by "Follow-up care" and "Environment" domains (4.69, each). Notably, "Transparency" and "Link between physician-patient" were the least rating domains (4.58 and 4.62, respectively), suggesting the timely provision of enhancement. ConclusionsWe propose "FETAL Surgery Global Satisfaction Tool" as a potential valuable instrument to assess patient-perceived satisfaction. It could provide insights into the quality of services offered by Fetal Medicine and Therapy Units. By identifying key areas for improvement, this tool could support continuous quality enhancement across Fetal Therapy programs globally. Further studies are warranted to evaluate its reliability and validity.

This study developed and validated the Childrens Medication Self-Efficacy Scale (CMSES), a psychometric tool grounded in Banduras self-efficacy theory, to assess medication self-efficacy in pediatric populations. Using data from 2,258 Chinese children, the four-item scale demonstrated robust psychometric properties through exploratory factor analysis (EFA) and confirmatory factor analysis (CFA), revealing a two-factor structure (magnitude and strength) that accounted for 76% of the total variance. The scale exhibited excellent internal consistency (Cronbachs = 0.92). Latent profile analysis (LPA) identified three distinct self-efficacy profiles and Receiver operating characteristic (ROC) curve analysis established a cut-off score to effectively distinguish low self-efficacy cases. These findings confirm the CMSES as a reliable and valid tool for evaluating medication self-efficacy in children, offering clinicians a practical instrument to design tailored interventions aimed at improving medication adherence in pediatric care.

BackgroundHandwriting is one of the fundamental skills for school-aged children, and it is an intricate and interacting process. An innovative computerized handwriting assessment system, the Smart Handwriting Analysis and Recognition Platform (SHARP) was used to provide a comprehensive and quantitative analysis of Chinese handwriting performance and identify children with handwriting difficulties objectively, accurately, and efficiently. Aims/ObjectivesThe purpose of this study is to provide quantitative data on childrens Chinese handwriting process and products using SHARP; to investigate the handwriting difficulties in children with neurodevelopmental problems, and to explore the associations between handwriting and sensorimotor performance, including fine motor skills, visual perceptual skills, visual-motor integration, and oculomotor proficiency. Material and MethodsTypically developing students and students with neurodevelopmental problems were recruited through teachers and therapists from six primary schools in Hong Kong. SHARP and tests of sensorimotor skills were performed. ResultsThe quantitative analysis using the SHARP system showed a progressive developmental change in handwriting performance among the typically developing children at their primary education. There were also significant differences between the typically developing students group and those with neurodevelopmental problems, both in handwriting speed and writing errors. Significant associations between the handwriting process parameters and fine motor skills were identified, while handwriting product parameters were closely correlated with visual-perceptual skills of these typical developing students. Conclusions and SignificanceSHARP appears to provide a comprehensive, quantitative and objective evaluation of students handwriting performances, thus assisting both teachers and parents in comparing students performance objectively, thus identifying students with handwriting difficulties as early as P1 level. It was observed that handwriting performance has strong relationships with young childrens visual perceptual skills and fine motor skills. The findings could help to facilitate early intervention for those students with handwriting problems.

Technology and its advancements have significantly impacted the quality of life for humans. With the progress in technology, screen dependency has also progressed in people of all ages. Despite available guidelines recent surveys indicate that the screen time in children is beyond the recommended levels. Excess screen time is known to create behavioural issues, delayed language development, poor expressive language, poor social skills, and traits like autism (aka virtual autism) in young children. Children with such traits demonstrate significant improvements in their social and language skills when screen time is reduced with adequate guidance and stimulation. The present study, thus, aimed to develop a parental education module to reduce screen time for preschool children with social communication deficits. At first, the contents of the module were developed through a parental in-depth interview conducted to find the issues faced by parents due to screen time in children, in addition to a focused literature search. A panel of experts rated the contents of the module in terms of relevance, age appropriateness and feasibility of implementation. Expert rating showed a high consensus (80%) among the raters and appropriateness of its contents. Furthermore, feedback received on the revised module revealed acceptance by the parents. The module is expected to improve the parental awareness, knowledge and attitude about the challenges of screen time in children and thereby control the effects of screen time in this group.

There exists a gap in existing patient education resources for children with chronic conditions. This pilot study assesses large language models (LLMs) capacity to deliver developmentally appropriate explanations of chronic conditions to pediatric patients. Two commonly used LLMs generated responses that accurately, appropriately, and effectively communicate complex medical information, making them a potentially valuable tool for enhancing patient understanding and engagement in clinical settings.

RationaleA scoping review was conducted to investigate knowledge gaps in the informatics research literature regarding sex differences in cognitive decline, identifying existing studies and areas where further studies are needed. Materials and MethodsWe searched Ovid and other databases for studies on sex differences and cognitive decline, focusing on publications in peer-reviewed informatics journals and conference proceedings from 2000 to 2023. The selected manuscripts were analyzed and summarized through discussion among three reviewers. ResultsA total of 13 articles were selected and examined for metadata and attributes analysis. Most studies are conducted in United States (n=5) and European Union (n=4), about a half are published after 2020 (n=6), and most studies are published in Springer and Elsevier. Our attributes-based analysis highlights the different aspects of reported studies such as task, method, dataset and its size, and sex-specific inferences. DiscussionSex-specific disparities in cognitive decline remain a critical issue in healthcare, yet most informatics research has primarily concentrated on identifying basic sex differences, such as tracking the progression of cognitive decline in men and women. While these studies are valuable, they fall short of addressing the more complex underlying causes of these sex-specific disparities in progression of cognitive decline. ConclusionThere is a significant gap using informatics in understanding how biological, social, and behavioral factors contribute to sex-specific disparities. This limited focus restricts the development of effective intervention strategies for mitigating sex-specific differences in cognitive health outcomes, underscoring the need for more comprehensive research that goes beyond mere identification to find the root cause of these disparities in healthcare.

BackgroundThe SARS-Cov-2 pandemic (COVID-19) sudden onset disrupted the direct access to face-to-face nutrition consultation fostering the rapid adoption of Telehealth by dietitians. ObjectiveThis study investigates Telehealth use among Lebanese Licensed Dietitians (LDs) amid COVID-19, in the absence of national Telehealth practical guidelines (TPG), and offers insights on Telehealth application under uncertainties of pandemic event, economic crisis, and destructed infrastructure occurring simultaneously in Lebanon. DesignA cross-sectional study during March 2023, using an anonymous online survey (44 questions), diffused through the Lebanese Order of Dietitians and social media platforms. ParticipantsThe sample (n=94) consisted mostly of females (98.9%) and young dietitians (mean(SD) age: 30.54(6.41) years) having a mean(SD) of 7.89(5.7) years of experience. Most respondents identified clinical nutrition as their primary practice area (87.2%), mostly in weight management (84%). Main outcome measuresTelehealth experience amid COVID-19 including tools utilized in remote consultations, barriers, facilitators, and perspectives of Telehealth use. Statistical analysesDescriptive analysis (counts, frequencies) using SPSS version 28. ResultsAlthough 48.4% of LDs reported using Telehealth prior COVID-19, this proportion increased to 97.8% during the pandemic. The most common tools used were WhatsApp (90.3%), Zoom (72.0%), and E-mails (41.9%). Reported barriers included bad internet connection (74.2%), patients preferring face-to-face consultation (61.3%), and patients lacking technical literacy (33.3%); benefits included scheduling and time flexibility (83.9%), decrease in practice-related costs (77.4%), and compliance with social distancing measures (53.8%). The majority agreed that Telehealth is needed (78.5%) and applicable in the Lebanese context (64.6%) and expressed the need for Telehealth trainings (78.5%) and TPG for nutrition care (74.2%). ConclusionsThis study recognizes increasing use of Telehealth in Lebanon, warranting the development of Telehealth nutrition care infrastructure comprising national regulations and evidence-based practical guidelines to respond to the innovation in the healthcare industry, and to assure Telehealth sustainability in LDs routine practice.

BackgroudGenerative language models (GLM) utilize machine learning algorithms to perform various tasks such as text generation, question response, and sentence completion by imitating the language that humans understand and use. PurposeThis study was to fine-tune the Llama2 language model using text data from books on the diagnosis and treatment of musculoskeletal system in physical therapy and compare it to the base model to determine its usability in medical fields. ResultsCompared to the base model, the fine-tuned model consistently generated answers specific to the musculoskeletal system diagnosis and treatment, demonstrating improved understanding of the specialized domain. ConclusionThe model fine-tuned for musculoskeletal diagnosis and treatment books provided more detailed information related to musculoskeletal topics, and the use of this fine-tuned model could be helpful in medical education and the acquisition of specialized knowledge.

BackgroundOne of the targets for the third sustainable development goals is to reduce worldwide maternal mortality ratio (MMR) to less than 70 deaths per 100,000 live births by 2030. To address issues affecting women and the newborns during childbirth and postnatal period, concerted efforts from governments and their stakeholders are crucial to maximize the use of technology to enhance frontline health professionals skills to provide the emergency obstetric and newborn care (EmONC). However, no study has garnered nurses and midwives perspectives regarding the application of technology-enhanced learning approach to provide on-job CPDs and factors that may influence the application of this training approach in the Rwandan context. MethodsThe study collected data from nurses and midwives from forty (40) public health facilities in remote areas nationwide. The study applied a qualitative descriptive design to explore and describe nurses and midwives perspectives on the feasibility and acceptability of technology enhanced learning approaches such as e-learning, phone-based remote training, and other online methods to provide trainings in EmONC. Two focus group discussions with EmONC mentor, two with nurses and midwives were conducted. Twelve key informant interviews were conducted. Participants were selected purposively. In total, 54 individuals were included in this study. A thematic approach was used to analyse data. ResultsNurses and midwives highlighted the need to provide refresher trainings about the management of pre-eclampsia. Most of the EmONC trainings are still provided face to face and the use of technology enhanced learning approaches have not yet been embraced in delivering EmONC CPDs for nurses and midwives in remote areas. Nurses and midwives found the first developed prototype of smartphone app training of the EmONC acceptable as it met the midwives expectations in terms of the knowledge and skills gap in EmONC. ConclusionAlthough the newly developed application was found acceptable, further research involving practical sessions by nurses and midwives using the developed application is needed to garner views about the ease of use of the application, relevance of the EmONC uploaded content on the app, and needed improvements on the app to address their needs in EmONC.

We assessed the risk of bias in randomized controlled trials (RCTs) of nutrition-only interventions and holistic frailty outcomes in older adults. We also explored associations between study-level factors and risk of bias. We searched Cochrane, PubMed, and Scopus for published trials between 01/01/2000 and 11/13/2024. Two persons independently screened each citation at the title and abstract, and full text, levels. They also independently conducted data extraction and assessed risk of bias using the Cochrane Risk of Bias 2 tool. We used responses on the tool to develop index scores between 0-1 for each included article, with higher scores indicating lower risk of bias. We regressed the index scores on four study-level factors, i.e., region of publication, year of publication, journal impact factor, and reported use of CONSORT guidelines. Fifteen articles were included in the study: three had low risk of bias, two had some concerns with bias, and ten had high risk of bias. Domain 2 on the Cochrane tool generated the most challenges with bias, largely due to poor reporting of intention-to-treat analysis and lack of information on how this issue might affect trial results. Median index scores were 0.52, 0.53, and 0.86 for articles with high, some concerns, and low risk of bias, respectively (p = 0.0479). However, the index scores were not associated with any study-level factors. Researchers in the field should note potential biases in the design and conduct of RCTs, especially in data analysis and - more specifically - intent-to-treat analysis.

IntroductionIt is estimated that large language models (LLMs) including ChatGPT is already widely used in academic paper writing. This study aims to investigate whether the usage of specific terminologies has increased, focusing on words and phrases frequently reported as overused by ChatGPT. MethodsA list of 142 potentially AI-influenced terms was curated from online discussions and recent literature documenting LLM vocabulary patterns, while 84 common academic terms in the medical field were used as controls. PubMed records from 2000 to 2024 were analyzed to track the frequency of these terms. Usage trends were normalized using a modified Z-score transformation. ResultsAmong the potentially AI-influenced terms, 100 displayed a meaningful increase (modified Z-score [&ge;] 3.5) in usage in 2024. The linear mixed-effects model showed a significant effect of potentially AI-influenced terms on usage frequency compared to common academic phrases (p < 0.001); the usage of potentially AI-influenced terms showed a noticeable increase starting in 2020. DiscussionThis study revealed that certain words, such as "delve," "underscore," "meticulous," "boast," and "commendable," have been used more frequently in medical and biological fields since the introduction of ChatGPT. The usage of these terms had already been increasing prior to ChatGPTs release, suggesting that ChatGPT accelerated the popularity of expressions already gaining traction. The identified terms can inform medical educators aiming to enhance awareness of language trends and promote best practices among trainees using LLMs.

Introduction.Digital technologies are increasingly influencing childrens lives, with many seeking digital platforms for nutritional education. This study aims to assess the usability and acceptability of Ajyal Salima, a nutrition awareness digital platform targeting children aged 9-11, in four Arab countries. Methods.A qualitative study was led across four countries: Lebanon, Bahrain, Palestine, and Jordan. Semi-structured focus groups were conducted with children and parents, and one-to-one interviews were held with teachers. Data was analyzed using thematic content analysis. Results.Four major themes emerged: platforms usability, content enjoyment, changes in childrens habits and recommendations to improve the platform. Overall, parents and teachers found the digital experience positive and useful and the content appropriate for children, particularly younger age groups. Challenges included registration difficulties, technical problems, internet accessibility, low parental involvement, and difficulties integrating the platform into teachers schedules. The platforms animations were less effective in sustaining childrens attention amid evolving digital standards. Conclusion.To enhance the platforms effectiveness, recommendations include simplifying the registration process, enhancing content interactivity, aligning the platform with school curricula, and equipping teachers with supportive resources. Fostering stronger school-family partnerships and engaging parents through community initiatives may be considered to maximize the platforms potential to promote healthier eating habits and improve nutritional awareness among children and their families, across the region.

IntroductionIsrael led a rapid vaccine rollout against COVID-19, leading to a local remission of the epidemic and rolling back of most public health measures. Further vaccination of 12-15-year-olds may be hindered by public perceptions of the necessity and safety of vaccination. Methodswe examined the considerations of vaccine hesitant parents (VHPs) regarding vaccination of children against COVID-19. The responses of 456 parents were surveyed and analyzed before FDA authorization of vaccination of children. Resultsparents who were vaccinated against COVID-19 were more likely to intend to vaccinate their children (r=-0.466, p<0.01). Low accessibility of vaccination may be a dissuading factor for VHPs more inclined to vaccinate. Vaccine efficacy and gaining a "Green Pass" were positively associated with an intention to vaccinate and statistically significant. VHPs inclined not to vaccinate indicated short development time and possible long term effects as dissuading factors. Discussionvaccine promotion should be tailored for VHPs positive and negative considerations for higher uptake.

PurposeThe purpose of this study was to validate Hospital at Home as an appropriate care option for patients of certain diagnostic-related groups and acuity levels. Patients and methodsWe compared outcomes for patients in a Hospital at Home program at Virtua Health in 2022 (N = 272) to traditional inpatients at Virtua hospitals during the same year who did not participate in the Hospital at Home program (N = 13879). We defined outcomes as recommendations for subacute rehabilitation (SAR) and final disposition upon inpatient discharge. Specifically, we searched electronic medical records for terms related to recommendation for SAR and discharge to SAR using text mining algorithms and a natural language processing (NLP) model to confirm these recommendations. ResultsWe observed that the proportion of patients in the Hospital at Home program that were recommended for SAR (0.147) was significantly different from the proportion of patients who remained in the hospital (0.361). Further, of those patients who received a recommendation, only 1 patient in the Hospital at Home group was discharged to SAR, while nearly half of those in the control group (proportion = 0.499) were transferred to SAR. ConclusionThe Hospital at Home program is a promising alternative to traditional inpatient care for patients with certain diagnoses and who meet certain clinical criteria.

BackgroundGenerative artificial intelligence (GAI) tools, such as large language models, have the potential to revolutionize medical research and writing, but their use also raises important ethical and practical concerns. This study examines the prevalence and content of GAI guidelines among Korean medical journals to assess the current landscape and inform future policy development. MethodsTop 100 Korean medical journals by H-index were surveyed. Author guidelines were collected and screened by a human author and AI chatbot to identify GAI-related content. Key components of GAI policies were extracted and compared across journals. Journal characteristics associated with GAI guideline adoption were also analyzed. ResultsOnly 18% of the surveyed journals had GAI guidelines, which is much lower than previously reported international journals. However, adoption rates increased over time, reaching 57.1% in the first quarter of 2024. Higher-impact journals were more likely to have GAI guidelines. All journals with GAI guidelines required authors to declare GAI use, and 94.4% prohibited AI authorship. Key policy components included emphasizing human responsibility (72.2%), discouraging AI-generated content (44.4%), and exempting basic AI tools (38.9%). ConclusionWhile GAI guideline adoption among Korean medical journals is lower than global trends, there is a clear increase in implementation over time. The key components of these guidelines align with international standards, but greater standardization and collaboration are needed to ensure responsible and ethical use of GAI in medical research and writing. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=106 SRC="FIGDIR/small/24303960v1_ufig1.gif" ALT="Figure 1"> View larger version (27K): org.highwire.dtl.DTLVardef@20d506org.highwire.dtl.DTLVardef@181a810org.highwire.dtl.DTLVardef@140cee8org.highwire.dtl.DTLVardef@1cff266_HPS_FORMAT_FIGEXP M_FIG C_FIG