Clinical Research Directory

Physician Response Evaluation With Contextual Insights vs. Standard Engines - Artificial Intelligence RAG vs LLM Clinical Decision Support

Clinical decision support tools powered by artificial intelligence (AI) are being rapidly integrated into medical practice. Two leading systems currently available to clinicians are OpenEvidence, which uses retrieval-augmented generation to access medical literature, and GPT-4, a large language model. While both tools show promise, their relative effectiveness in supporting clinical decision-making has not been directly compared. This study aims to evaluate how these tools influence diagnostic reasoning and management decisions among internal medicine physicians.

Benchmarking Large Language Models Against Tumour Boards for Oncology Treatment Recommendations

BEACON (Benchmarking AI for Clinical Oncology decisioNmaking) is a prospective, multicentre, comparative, blinded, non-interventional benchmark evaluating the treatment recommendations of five frontier large language models (LLMs) against the recommendations of multidisciplinary tumour boards (RCP) in oncology treatment planning. One hundred standardised synthetic cases (20 per localisation, across breast, lung, urological, digestive and gynaecological cancers) are submitted as identical structured input to two independent tumour boards per localisation and to five frontier LLMs. Each recommendation - human or model - is decomposed into five predefined decision domains (intent, surgery, radiotherapy, systemic therapy, work-up and biomarkers) and scored 0/1/2 for concordance against a two-tier reference: the consensus of the two tumour boards, complemented by an a priori locked guideline matrix (ESMO, NCCN). The primary endpoint is domain-level concordance between LLM and RCP consensus, expressed as a linearly weighted Cohen's kappa. A co-primary safety endpoint captures the proportion of recommendations carrying serious harm potential, because concordance alone can conceal dangerous errors. Because expert boards may disagree with one another on identical cases, model performance is always interpreted against the human consensus. BEACON is designed as reusable, openly licensed, pre-registered infrastructure: all synthetic cases, evaluation rubrics, the locked guideline matrix, scoring algorithms and verbatim prompts are released for full reproducibility.

Preliminary Evaluation of a Large Language Model-Based Tool for Complex Surgical Decision Support in Lung Cancer

This study is an exploratory effect-size estimation study, with the following specific objectives: ① to estimate the point estimate and 95% confidence interval of the Win Ratio for the experimental group (GAPS-Agent) versus the control group (large language model) in blinded pairwise preference judgments by thoracic surgery expert adjudicators, to serve as a sample size planning parameter for subsequent multicenter confirmatory clinical trials; ② to preliminarily evaluate the value of GAPS-Agent within clinical workflows.The hypothesis of this study is as follows: compared with a general-purpose large language model without medical enhancement (control group), a structured agentic workflow optimized on the basis of the GAPS evaluation framework (GAPS-Agent, experimental group) can help junior resident physicians generate clinical decision plans for complex lung cancer cases that are more strongly preferred by senior thoracic surgery expert adjudicators.

Improving AI-Assisted Medical Diagnosis and Triage by the General Public

This study is a randomized controlled trial (RCT) investigating whether access to a new LLM interface can improve medical triage and diagnostic accuracy for laypeople compared to access to a standard LLM interface. It addresses previous findings where laypeople using standard LLMs performed worse than those using conventional methods (e.g., web search) due to incomplete symptom sharing and poor interpretation of AI advice. To address this, the research tests a structured LLM system that proactively asks clinical history questions before providing a standardized, easy-to-read diagnostic output.

A Simulated Case Study of a Peritoneal Dialysis-Specialized Large Language Model Assisting Doctors in Improving Decision-Making in Peritoneal Dialysis Management

This study is a randomized controlled trial based on simulated clinical cases, aiming to establish a standardized evaluation system for PD physicians, to assess the differences in PD management quality between a workflow assisted by a PD-specialized large language model and physician-only decision-making, and to identify potential risks (such as generating obviously erroneous or even harmful recommendations). This simulated clinical case framework not only supports standardized and blinded evaluation, but also provides preliminary evidence for the model's effectiveness and safety before its deployment in real clinical settings, while avoiding direct impact on real patients.

Evaluation of an Artificial Intelligence-enabled Clinical Assistant to Support Thyroid Cancer Management

This study aims to evaluate the clinical feasibility of adopting artificial intelligence (AI)-based models to improve clinical management of thyroid cancer.

A Multimodal AI Agent for Ophthalmic Clinical Decision Support

This study is a multicenter randomized controlled trial evaluating the effectiveness and safety of EyeAgent, a multimodal artificial intelligence (AI) agent designed to assist ophthalmologists in clinical decision-making. Participants will be recruited from ophthalmology clinics and hospitals in Hong Kong and mainland China. The AI agent acts as a digital co-pilot, analyzing patient images and clinical history to provide diagnostic and management recommendations. The trial aims to determine whether the use of the AI agent improves diagnostic accuracy, treatment decision-making performance, report generation, workflow efficiency, and user satisfaction compared to standard clinical practice.

Acute Myocardial Infarction Clinical Intelligent Decision Support System

Acute Myocardial Infarction (AMI) remains the leading cause of cardiovascular mortality globally. In China, while the incidence of AMI is escalating at an annual rate of 5.2%, significant clinical challenges persist: diagnostic delays in primary care facilities exceed 40%, and the "Door-to-Balloon" (D2B) compliance rate in tertiary hospitals stagnates at a mere 65%. These figures underscore systemic deficiencies, including inefficient emergency response, regional resource disparities, and fragmented longitudinal care. Although Large Language Models (LLMs) provide a transformative technical foundation for AMI management, their clinical translation is hindered by critical bottlenecks, such as non-standardized data interfaces, limited model interpretability, inadequate hardware infrastructure at the grassroots level, and the inherent tension between data privacy and training requirements. This research proposes a comprehensive implementation strategy for an AI-driven intelligent decision-making system for AMI. On a theoretical level, the study establishes a tripartite framework of "Technological Adaptation, Scenario Implementation, and Safeguard Mechanisms." By introducing a data governance scheme based on federated learning and multimodal fusion, and constructing a "Technical-Clinical-Economic" multidimensional evaluation model, this work bridges the theoretical divide between advanced technology and clinical practice. On a practical level, the study develops adaptive gateways and lightweight models to facilitate pervasive deployment in resource-constrained settings, optimizes the full-cycle clinical workflow to improve patient outcomes, and provides a scalable, replicable pathway for implementation. Focusing on four core challenges-technological compatibility, clinical workflow integration, the balance between privacy and performance, and the establishment of scientific evaluation systems-this research aims to surmount existing translation barriers. It seeks to enhance the quality and efficiency of AMI care while providing a seminal reference for the clinical transformation of AI in other medical specialties.

Effect of Perception-based Interventions on Public Acceptance of Using Large Language Models in Medicine

Large language models (LLMs) show promise in medicine, but concerns about their accuracy, coherence, transparency, and ethics remain. To date, public perceptions on using LLMs in medicine and whether they play a role in the acceptability of health care applications of LLMs are not yet fully understood. This study aims to investigate public perceptions on using LLMs in medicine and if interventions for perceptions affect the acceptability of health care applications of LLMs.