Jul 2026· International Journal of Surgery· 0 citations· 8 references
TL;DR
This paper conducts a comprehensive analysis of evaluation methods, deployment processes, and governance strategies for LLMs in the healthcare field, focusing on three key issues: model version drift, multilingual external validation, and prompt injection security governance.
Abstract
Large language models (LLMs) are gradually evolving from text generators into cognitive interfaces and decision-making infrastructure that can be integrated into clinical systems. Although LLMs have demonstrated tremendous potential in retrieving evidence, structuring medical records, and facilitating clinician-patient communication, their clinical translation still faces systemic challenges in the highly heterogeneous, high-risk, and strictly regulated field of healthcare. This paper aims to utilize bibliometric methods to analyze the current global research landscape and trends regarding LLMs in clinical and healthcare scenarios. Furthermore, it proposes a systematic framework for future research directions, with the goal of providing more actionable guidance for subsequent research and clinical translation pathways in this field. Combining the results of global informatics analysis with relevant research advances, this paper conducts a comprehensive analysis of evaluation methods, deployment processes, and governance strategies for LLMs in the healthcare field, focusing on three key issues: model version drift, multilingual external validation, and prompt injection security governance. Future research on LLMs in the healthcare field should shift its focus from merely demonstrating model performance to building a closed-loop system of clinical evidence and governance that is reproducible, monitorable, auditable, and explainable. Through model drift governance, multilingual external validation, and system-level security governance, it is possible to enhance the models’ long-term reliability, generalization fairness, and the controllability and accountability of their deployment, respectively. On this basis, LLM applications may be cautiously expanded from low-risk, reviewable tasks to medical education, clinical decision support, and real-world healthcare workflows.
The integration of Large Language Models (LLMs) into healthcare is poised to revolutionize various aspects of medical practice, including clinical decision‐making, patient care, and medical research. This review explores the applications of LLMs such as ChatGPT‐3, ChatGPT‐4, and BERT in healthcare, focusing on their potential to enhance disease diagnosis, treatment planning, and personalized care. The paper presents a comprehensive bibliometric analysis of the growing body of research, highlighting key trends, influential authors, institutions, and geographical contributions. Despite their promise, significant challenges remain, including model accuracy, data privacy, ethical concerns, and the need for domain‐specific fine‐tuning. This review examines the moral and technical challenges associated with deploying LLMs in healthcare, including biases, a lack of transparency, and issues related to model interpretability. The paper further emphasizes the importance of robust frameworks for ensuring ethical usage. It proposes future research directions to address these challenges, including the development of specialized healthcare models, enhanced transparency, and improved integration into clinical workflows. Ultimately, this review aims to inform healthcare professionals, researchers, and policymakers about the transformative potential of LLMs in healthcare while underscoring the critical issues that must be overcome for their widespread adoption.
This article is categorized under:
Application Areas > Health Care
Fundamental Concepts of Data and Knowledge > Big Data Mining
Technologies > Artificial Intelligence
Md Belal Bin Heyat, A. Rehman, H. M. Zeeshan et al.· WIREs Data Mining and Knowle...· 0 citations
Current evidence indicates that LLMs have substantial potential to enhance healthcare delivery, research, and personalized medicine, but they should currently be regarded as supportive tools rather than autonomous clinical decision-makers.
Antoni Klamka, Paulina Kawalec, Kamil Bronikowski et al.· 0 citations
This review provides a structured overview of recent progress in Med-LLMs by examining their major application areas, key challenges, and emerging future directions.
Yuyang Sha, Li Yu, Zejia Lin et al.· Military Medical Research· 0 citations
A thorough review of the developments in LLM technologies, their uses in clinical and administrative settings, as well as their ethical considerations are reviewed to suggest a conceptual structure for responsible implementation that will ensure both technological innovation and patient safety, as well as regulatory compliance and ethical health care practices.
Noah Wright· International Journal of Mod...· 0 citations
A scoping review of 24 PubMed-indexed studies published between 2023 and 2026 was conducted to assess current applications, benefits, limitations, and future directions of LLMs in healthcare.
Antoni Klamka, Paulina Kawalec, Kamil Bronikowski et al.· Quality in Sport· 0 citations
With the evolution of intelligent systems, there is an increasing need to incorporate multimodal inputs, such as medical images, clinical text, structured records, and physiological signals, to facilitate precise and contextually informed clinical decision-making. To cope with multimodal healthcare data, challenges like computational complexity, data heterogeneity, and privacy concerns are rising regardless of advancements in large language model (LLM) capabilities. Applying the PRISMA methodology, this systematic review examined the present status of research on multimodal large language models (MLLMs) in healthcare, published between 2019 and 2024. A total of 164 studies were reviewed, focusing on the types of multimodal data utilized, the fusion techniques and LLM architectures applied, together with the key challenges and emerging opportunities in multimodal processing. The analysis reveals that many existing approaches still lack a comprehensive understanding of MLLM architectures and their role in clinical decision-making, and they also often lack robust privacy-preserving mechanisms and sufficient validation in real-world clinical workflows. Based on these observations, we recommend a customizable framework consisting of interoperable components that can be configured to meet the needs of specific clinical use cases. Instead of proposing a new algorithm or rigid pipeline, this recommendation incorporates a retrieval-augmented generation (RAG) framework coupled with an LLM that can integrate multimodal healthcare data. A central focus of this design is the integration of privacy-preserving retrieval strategies with RAG workflows to ensure secure handling of multimodal data, which highlights the need to examine the role of multimodal data encoding and storage, encrypted retrieval, reranking of retrieved data, multimodal prompt engineering, and secure content generation. Together, these components enable the integration of RAG with MLLMs in a way that supports privacy-aware, explainable, and clinically grounded decision support. The insights presented aim to guide future efforts to develop secure, explainable, and clinically integrated intelligent systems that take advantage of RAG and multimodal LLM.