[Paper Review] From RAGs to riches: Utilizing large language models to write documents for clinical trials

Nigel Markey, Ilyass El-Mansouri+3 more|arXiv (Cornell University)|Feb 26, 2024

Biomedical Text Mining and Ontologies6 citations

TL;DR

The paper investigates using retrieval-augmented generation and large language models to write documents for clinical trials.

ABSTRACT

This manuscript has now been published: - Link to article on journal website: https://journals.sagepub.com/doi/10.1177/17407745251320806 - Pubmed link: https://pubmed.ncbi.nlm.nih.gov/40013826/

Motivation & Objective

• Motivate the use of LLMs and RAGs to automate clinical trial document authorship.
• Assess how RAGs can support drafting and editing of trial-related texts.
• Highlight potential benefits and limitations of LLM-based document generation for clinical trials.

Proposed method

• Proposes an approach using retrieval-augmented generation with large language models to draft clinical trial documents.
• Discusses integration of retrieval mechanisms with LLMs to generate structured trial documents.
• Illustrates the workflow and considerations for applying LLMs in clinical trial writing.

Experimental results

Research questions

• RQ1 Can RAG-enabled LLMs effectively draft clinical trial documents?
• RQ2 What are the practical considerations and limitations of using LLMs for trial writing?
• RQ3 What workflow and quality controls are needed when deploying LLMs for clinical documentation?

Key findings

• The manuscript presents a method combining RAGs with LLMs to write clinical trial documents.
• The paper is six pages long with two figures.
• The work has been published in Clinical Trials: Journal of the Society for Clinical Trials (2025).