BACKGROUND:The effective translation of scientific insights into public health action remains a challenge, despite ongoing efforts to address obesity. OBJECTIVE:To apply an infodemiological framework and examine trends in public and scientific interest from 2004 to 2023 in obesity and two behavioral and sociocultural proxy topics, the Mediterranean diet (MedDiet) and body positivity. METHODS:Data sources included Google Trends, Wikipedia page views, and SCOPUS publications. Pearson's correlation assessed linear associations. Univariate and multivariate forecasting models projected future trends. RESULTS:Scientific interest in obesity moved in parallel with MedDiet (r = 0.96) and body positivity (r = 0.89), whereas public interest measured via Google Trends were inversely associated with these topics (r = -0.62; r = -0.61). While public interest in MedDiet and body positivity aligned with scientific interest, interest in obesity per se declined over time (r = -0.89). Univariate forecasting suggests public interest in obesity will remain low through 2030, while exploratory multivariate analysis indicates a possible rebound, coinciding with a projected decline in body positivity interest. CONCLUSION:These findings reveal a divergence between scientific attention and public engagement patterns with obesity over time. While declining searches may partly reflect evolving terminology, digital trend analysis offers valuable tools for identifying gaps between research focus and public attention to support responsive, evidence-informed public health communication.
BACKGROUND:Healthcare users can be assisted by a medicine or drug recommendation system (MRS) by understanding their needs and supporting informed decisions based on complex knowledge. The biggest challenge is analyzing users' sentiments, as human language characterizes user-generated data in many complex ways. METHODS:This paper proposes a sentiment analysis (SA)-based drug recommender system using an optimized deep learning (DL) model with an enhanced transformer-based feature representation mechanism. The proposed system comprises four stages: preprocessing, feature extraction, classification, and medicine recommendation. Firstly, drug reviews collected from public repositories are pre-processed to improve data quality. In the feature extraction stage, the pre-processed data are fed into the Multi-Head Attention Bidirectional Encoder Representations from Transformers (MHABERT) model to capture deep semantic and contextual features. The user sentiments are then classified using Golden Jackal Optimized Bidirectional Long Short-Term Memory (GOBLSTM). Finally, medicine recommendations are generated by comparing the similarities of drug reviews. RESULTS:The Kaggle medicine recommendation and Yelp health datasets are used to evaluate the proposed method. Experimental results demonstrate maximum accuracies of 98.43% and 98.30% on the Kaggle and Yelp datasets, respectively, outperforming existing approaches. CONCLUSION:The combination of SA and drug recommendation enhances patient outcomes, optimizes medicine selection, and reduces adverse reactions.
Continued use of Electronic Health Records (EHRs) remains a significant challenge despite widespread digitalization. Existing models often overlook key psychological and informational factors essential for continued engagement. This study extends the Unified Theory of Acceptance and Use of Technology (UTAUT) by incorporating digital health literacy, trust, and privacy concerns to better explain continued EHR use. A cross-sectional survey across 21 Turkish cities yielded 480 valid responses from users of the national EHR system (E-nabız). Data were analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM) to identify usage intention predictors, and Necessary Condition Analysis (NCA) to determine must-have factors for continued use. UTAUT constructs were confirmed as relevant predictors of continued EHR use. Effort expectancy, facilitating conditions, and trust emerged as necessary conditions. Trust also reduced the impact of privacy concerns, which had only a weak negative effect. Social influence affected intentions but was not a necessary condition. Digital health literacy significantly predicted performance expectancy, effort expectancy, and facilitating conditions and proves critical to ongoing EHR usage. Educational interventions aimed at enhancing digital health literacy can significantly increase EHR continued use through multiple pathways.
OBJECTIVES:This study conducted an informatics system evaluation of two LLMs (GPT-4o and DeepSeek-V3) for patient education, combining clinician-rated quality with patient-perceived usability across thematically stratified queries. MATERIALS AND METHODS:In a blinded, within-subject design, 16 frequently asked questions about biologic therapies were categorized into three domains: treatment/drug selection, safety/adverse effects, and special conditions/daily life. Responses were standardized, generated without external retrieval, anonymized as A/B pairs. Thirty physicians assessed clinical appropriateness, scientific accuracy, comprehensiveness, while 60 patients rated readability, understandability, actionability, perceived adequacy, decision support, and trust on 5-point Likert scales. Analyses included paired t-tests, Holm/FDR corrections and two one-sided tests (TOST) to distinguish statistical non-difference from practical equivalence. RESULTS:Physicians rated GPT higher across all domains (p < .002), with largest gaps in safety/side effects and treatment/drug selection. Patients favored GPT for understandability, actionability, and decision support (p < .001), while readability, adequacy, trust, and reading time were statistically and clinically equivalent. CONCLUSION:Findings highlight the need for topic-aware governance: guideline-dense queries suited to retrieval-augmented generation and checklist compliance, and context-sensitive queries requiring uncertainty signaling and human oversight. This layered approach advances health informatics by defining where LLMs may substitute versus where they require verification, supporting safe and auditable integration into patient education.
To identify research topics in medical chatbots and analyze their temporal trends, geographic distributions, and journal preferences. Latent Dirichlet Allocation (LDA) topic modeling was applied to 9,650 publications (1986-2024), extracting eight core topics integrated with time-series analysis, geographic statistics, and journal associations. Eight topics were identified. Temporal trends revealed three phases: technology incubation (2000-2015), rapid breakthrough (2015-2020), and application consolidation (2020-2024). Geographically, the United States, China, and the United Kingdom dominated research output (46.0%). Journal analysis highlighted Journal of Medical Internet Research (JMIR) (7.1%), Journal of the American Medical Informatics Association (JAMIA) (5.4%), and IEEE Journal of Biomedical and Health Informatics (IEEE JBHI) (3.2%) as top contributors, with JMIR and JAMIA reinforcing clinical informatics and digital therapeutics. Research on medical chatbots needs to balance technical feasibility and clinical value. Future research should focus on three directions: developing validation frameworks for leveraging large language models in clinical applications (LLMs), establishing transnational data-sharing infrastructure, and creating ethical governance mechanisms that ensure responsible innovation while maintaining health equity.
