DIGITAL HEALTH最新文献

Background: Large language models (LLMs) are increasingly used in healthcare for patient education and clinical decision support. However, systematic benchmarking in real-world clinical contexts remains limited, particularly for high-risk conditions such as hip fractures.

Objective: To evaluate and compare the performance of three state-of-the-art LLMs-DeepSeek-V3-FW, Gemini 2.0 Flash, and ChatGPT-4.5-in answering standardized patient questions on hip fracture management.

Methods: Thirty standardized questions covering general knowledge, diagnosis, treatment, and rehabilitation were developed by three specialists in orthopedics and traumatology. Each LLM generated responses independently. Three experienced orthopedic surgeons assessed accuracy (4-point scale) and comprehensiveness (5-point scale). Statistical analyses included Kruskal-Wallis and chi-squared tests.

Results: All models demonstrated high reliability, with 96.7% of responses rated "Good" or "Excellent" and none rated "Poor." Mean accuracy scores were comparable across models, and comprehensiveness averaged 4.8/5. DeepSeek-V3-FW tended to provide longer, structured answers and performed best in general knowledge, while Gemini 2.0 Flash excelled in diagnosis and rehabilitation and produced the most concise responses. ChatGPT-4.5 offered shorter, conversational answers with similar accuracy and detail.

Conclusions: The three LLMs showed strong capabilities in delivering accurate and comprehensive information on hip fracture care, highlighting their potential as tools for patient education and clinical support. Differences in style and domain-specific strengths suggest complementary roles. Further research is needed to validate safety and integration into clinical workflows.

Objective: To systematically evaluate the effect of digital intervention on improving routine vaccination in the elderly and to conduct a comparative analysis of different intervention modalities using network meta-analysis (NMA).

Methods: PubMed, Web of Science, The Cochrane Library, Embase, Scopus, CINAHL, and WanFang Data were searched for randomized controlled trials (RCTs) using digital interventions to promote vaccination in older populations from inception to 15 June 2024. We performed a final update of the literature search in May 2025; no additional eligible studies were identified. Two researchers independently screened the literature, extracted data, and assessed the risk of bias in the included studies, and an NMA was performed using RevMan 5.4 and R Studio, PROSPERO Registration Number: CRD42024527483.

Results: Eleven RCTs were included. The traditional meta-analysis demonstrated a small but statistically significant increase in influenza vaccination rates (RR = 1.01, 95% CI [1.01, 1.01], P < 0.00001), accompanied by substantial heterogeneity (I ² = 86%). Pneumococcal vaccine uptake was significantly enhanced (RR = 1.11, 95% CI [1.03, 1.18], P < 0.01), with moderate heterogeneity (I ² = 46%). The single study on the herpes zoster vaccine reported a statistically significant effect, whereas COVID-19 vaccine reminder interventions showed no significant efficacy. In the NMA, video-based interventions ranked first based on the surface under the cumulative ranking curve, but all pairwise comparisons between different intervention modes crossed the null value.

Conclusion: Digital interventions show a significant, yet highly heterogeneous, positive impact on vaccination rates in older adults. While video-based education showed the highest ranking probability, the current evidence is insufficient to conclude that any specific digital modality is statistically superior to others. Due to the limited included studies, the findings need to be supplemented by more high-quality studies. Future research should focus on newer digital technologies to help the older population keep up with the "digital intelligence era."

Objectives: Visually-induced motion sickness (VIMS) remains an unsolved issue when using head-mounted displays (HMD) in immersive virtual reality (VR) systems. We investigate the influence of restricted visual field on sensorimotor incongruence of postural control in an immersive VR environment, as a surrogate marker of VIMS while standing steadily and viewing ambiguous optical-flow visual stimuli.

Methods: Twenty-seven healthy participants wore HMD and viewed optical flow visual stimuli (called the plaid motion pattern) using a VR system. Visual stimuli were presented on a full-screen display, either only in the center or periphery of the visual field, or nowhere (Natural). To evaluate standing posture stability, we measured spatial (length and area of the center of the pressure (COP)) and temporal dynamics of body sway using a stabilometer. Subjective feelings of VIMS were assessed for each visual stimulus condition.

Results: The full-screen condition significantly worsened the COP measurements and feelings. The COP areas of standing balance under the center of the visual field condition were significantly smaller than those under the full-screen condition; however, the periphery of the visual field condition was comparable to the full-screen condition. The condition effects for the conditions (natural, center of the visual field, periphery of the visual field and full-screen) were observed in the COP measurements and subjective feelings.

Conclusion: Optical-flow visual stimuli can induce body sway, suggesting sensorimotor incongruence in postural control. Ancillary findings demonstrated that subsequent to full-screen presentation, the periphery of the visual field potentially contributes to spatial and temporal dynamics of body sway and the sensorimotor incongruence rather than the central area.

Introduction: Frailty is defined as a clinically recognised state of increased vulnerability, reflecting a decline in an individual's psychological and physical reserves. Digitally-enabled interventions are increasingly utilised to monitor and support the health of older adults. Evidence on the effectiveness of digitally-enabled interventions in reducing frailty is limited. This systematic review aimed to investigate the types of digitally-enabled interventions tested, with what goals with respect to frailty, and the resulting outcomes.

Method: Medline, CINAHL, Scopus, PsychInfo and Embase were searched from time of origin until July 2024. Peer-reviewed RCTs assessing the impact of digitally-enabled interventions on older adults were included. Outcome measures explored were frailty, cognitive status, mental health, quality of life, adherence and usability. Data was extracted independently by two people using Covidence platform. Narrative synthesis was performed for all studies and meta-analysis was performed for outcomes reported in four or more studies.

Results: From 4476 titles and abstracts screened, 17 studies were included following full text review. Overall, 12 studies included exercises as a component or the sole form of intervention. The mean duration of intervention was 4.04(SD2.56) months. Mean adherence to the intervention was 59% which was lower in exercise-based intervention. The most and least reported frailty-specific outcome was walking speed (n = 8) and self-reported exhaustion level (n = 2). Meta-analysis showed non-exercise-based interventions showed significant improvements in SPPB. There was no statistically significant change in Timed-up and Go and handgrip strength. Narrative synthesis indicates there was insufficient evidence to evaluate the impact of digital interventions on frailty, frailty-specific outcomes, mental health, activities of daily living, health-related quality of life, sleep and cognition.

Conclusion: The findings suggest low technological readiness and adherence among digitally-enabled interventions for older adults. Narrative synthesis of overall frailty and outcome measures showed mixed results and limited evidence on the impact of digital interventions on frailty and outcomes.

Background: As part of the global digital health transformation and Saudi Vision 2030 priorities in chronic disease prevention, understanding public acceptance and preferences for telenutrition services is crucial for developing effective digital nutrition care platforms.

Objectives: This study aimed to adapt, translate, and validate a U.S.-based national telemedicine survey for the Saudi context, to assess adults' prior use, comfort, willingness, and preferences regarding direct-to-consumer (DTC) telenutrition services.

Methods: A cross-sectional survey was completed by 930 Saudi adults recruited from major urban cities. The survey instrument was adapted, translated, and validated using standard procedures, including internal consistency reliability using Cronbach's alpha coefficient. Descriptive statistics, chi-square tests, binary logistic regression, and two-step cluster analysis.

Results: Among 930 respondents, 61.5% expressed comfort and 54.9% reported willingness to use telenutrition services, increasing to 78.5% when referring to family members. Key predictors of willingness included previous experience with virtual care (OR = 1.52, p < 0.001), comfort with video consultations (OR = 1.39, p < 0.001), and access to personal health records (OR = 1.20, p < 0.041). Willingness was significantly higher among younger adults (18-29 years), employed individuals, and residents of Jeddah. Cluster analysis revealed three segments: digitally engaged and receptive, selective and independent, and supportive but personally hesitant-each characterized by differing levels of comfort and willingness to use DTC telenutrition services. Notably, only 1.3% could identify a telenutrition platform in Saudi Arabia.

Conclusions: These findings underscore the need for culturally tailored, digitally inclusive telenutrition platforms that address sociodemographic differences. Investing in digital health literacy and infrastructure is essential for empowering nutrition care delivery in Saudi Arabia.

Background: Children's mental health issues are widespread in the United Kingdom, yet access to traditional services remains limited. Digital mental health interventions (DMHIs) offer a scalable alternative by reducing barriers such as cost, stigma, and location. However, most DMHIs target either children or parents individually, overlooking the potential of combined approaches.

Objective: This review assessed the effectiveness of DMHIs that include both parent- and child-delivered components for children aged 12 and under. The review also examined user experiences and compared DMHIs with traditional mental health support services.

Methods: A systematic review was conducted following PRISMA guidelines, searching six electronic databases. Studies were included if they evaluated DMHIs involving both children and parents/guardians, focused on mental health outcomes, and were peer-reviewed. Data extraction and risk-of-bias assessments were performed independently by reviewers.

Results: Of 3703 screened studies, three randomised controlled trials (n = 630) met inclusion criteria. These showed improvements in child anxiety, social competence, and parental nurturing behaviours. Parents reported positive engagement, citing flexibility and ease of use. However, therapists raised concerns about long-term implementation. Due to study heterogeneity and small sample sizes, a meta-analysis was not feasible; a narrative synthesis was conducted.

Conclusions: DMHIs that incorporate both parent and child components are comparatively rare but show promise as a valid means of enhancing child and parent well-being. However, further research is needed to refine these interventions, address barriers to engagement, and optimise their broader application, particularly in community settings, to help reduce the pressure on clinical services.

Objective: This study aimed to evaluate the quality, reliability, and readability of online patient-centered information related to the management of gummy smile.

Methods: A systematic search was conducted using Google, Yahoo, and Bing to identify websites providing patient-oriented information on gummy smile treatments. A total of 257 websites met the inclusion criteria and were analyzed. Content quality was assessed using the DISCERN instrument, Journal of the American Medical Association (JAMA) benchmarks, and the Health on the Net (HON) code certification. Readability was evaluated using the Flesch Reading Ease Score (FRES), Flesch -Kincaid Grade Level (FKGL), Simplified Measure of Gobbledygook (SMOG) index, and Coleman -Liau index.

Results: The overall quality of online information was low to moderate, with a mean DISCERN score of 40 ± 9.9. Only 11 websites were certified by the HON code, indicating limited adherence to established standards for trustworthy health information. According to JAMA benchmarks, only two websites fulfilled all four criteria. Readability analysis demonstrated that the content was relatively complex, with a mean FRES of 60.1 ± 9.2 and a mean FKGL of 8.9 ± 1.8, exceeding the recommended reading level for the general public.

Conclusions: Online patient-centered information regarding gummy smile is generally of suboptimal quality and readability. The limited number of reliable and easily understandable resources underscores the need for improved quality control, standardization, and patient-focused content development. Enhancing the accessibility and reliability of online information may support better patient understanding and informed decision-making in dental aesthetics.

Background: FysBot is a ChatGPT-based mobile app developed to promote physical activity among adults living with obesity. This pilot study aimed to evaluate the feasibility and usability of FysBot.

Methods: A 6-week single-arm pilot study was conducted in which patients from an obesity rehabilitation clinic in Norway used FysBot. This pilot study employed an explanatory sequential mixed-methods design combining questionnaires and post-intervention interviews. Participants completed questionnaires at baseline and weeks 2, 4, and 6, assessing leisure-time physical activity (Godin Leisure-Time Exercise Questionnaire (GODIN)), motivation (Behavioral Regulation in Exercise Questionnaire-2 and relative autonomy index (RAI)), Self-Efficacy for Exercise (SEE), and System Usability Scale (SUS). Semi-structured interviews were conducted to explore user experiences further. Quantitative data were analyzed descriptively, with multiple imputations for missing data, while qualitative data were analyzed thematically.

Results: Fifty-three participants were eligible, 36 completed baseline, and 17 completed the final follow-up. App engagement declined steadily, with most participants ceasing use after week 2. The mean SUS score was 51.3, indicating below-average usability. The median of self-reported leisure-time physical activity (GODIN: 34-40) and overall motivation (RAI: 8.3-9.8) showed small, non-significant increases, while identified regulation increased significantly (2.8-3.3; p = 0.04) and SEE decreased (58-49). Qualitative findings supported these results, showing that participants valued the chatbot's motivational potential but experienced technical problems and limited personalization.

Conclusions: This study offers insight into the potential of a ChatGPT-based physical activity app for adults living with obesity and highlights key areas for refinement. Future iterations should incorporate user-requested features through iterative co-design, with enhanced personalization and guidance to improve relevance and engagement.

Background: Low birth weight (LBW) is a leading cause of death for newborns and increases chronic disease risks later in life. Early identification of LBW risk is crucial.

Aim: The objective of this study was to develop predictive models for LBW using boosting ensemble machine learning, with a focus on features available during early pregnancy, such as pre-pregnancy body mass index, body height, and blood pressure before 20 weeks of pregnancy.

Methods: This is a retrospective cohort study. We used electronic medical records in four hospitals in Taiwan where pregnant women received prenatal care from January 2016 to July 2019, including 6719 pregnant women. Data preprocessing involved normalization, one-hot encoding, and a synthetic minority oversampling technique for class imbalance. Boosting ensemble methods were used to build the LBW predictive models.

Results: The mean diastolic blood pressure (DBP) in early pregnancy (<20 weeks) was 66.5 mmHg, 29.6% had experienced abortion, 8.7% delivered LBW, 12.2% were overweight or obese before pregnancy, and 18.3% had elevated or stage I hypertension before 20 weeks of pregnancy. Lightweight Gradient Boosting Machine was the best-performing LBW model, with an area under curve of 0.96 and an accuracy of 93.4%. Early pregnancy DBP, maternal height, and number of abortions were the most important features.

Conclusions: The LBW prediction model performed well. Nurses could use the model to assess LBW risk and intervene early. Preventive efforts could be directed to blood pressure management starting early pregnancy, nutritional support for short mothers, and self-care for women with a history of abortions.

Objective: To develop an attention-based convolutional neural network (CNN) pipeline for early prediction of pathological complete response (pCR) to neoadjuvant chemotherapy (NAC) in breast cancer, improving feature selection and interpretability in whole slide image (WSI) analysis.

Methods: A retrospective analysis was conducted on 384,076 tiles extracted from 122 Hematoxylin and Eosin-stained WSIs, divided among an investigational cohort (IC, 82 patients enrolled at IRCCS Istituto Tumori "Giovanni Paolo II"), a validation cohort (VC, 20 patients, same Institution), and an external validation cohort (EVC, 20 patients belonging the Yale trastuzumab response cohort public dataset). WSIs were first annotated and then automatically segmented into tiles. Noninformative regions were filtered using Mini-Batch C-Fuzzy K-Means. Remaining tiles were analyzed using a CNN with a Convolutional Block Attention Module, prioritizing both histological features and tiles critical for predicting pCR.

Results: The model achieved robust performance: 81.4% AUC, 81.3% accuracy, 80.0% specificity, and 83.3% sensitivity in IC; 80.9% AUC, 80.0% accuracy, 85.78% specificity, and 66.7% sensitivity in VC; and 76.2% AUC, 70.0% accuracy, 71.4% specificity, and 66.7% sensitivity in EVC. The EVC, consisting of WSIs at 20× magnification compared to the 40× magnification of IC and VC, demonstrated the model's robustness to varying resolutions.

Conclusion: This is an innovative pipeline that not only improves prediction but also enhances the clinical utility of digital pathology.