Gen AI-Powered Evidence Generation: Implementing Advanced RAG Architecture for Sensitive Data in HEOR Applications

Moderator

Nicola Waddell, Pharmacoevidence Pvt. Ltd., London, United Kingdom

Speakers

Karam Diaby, BSc, MSc, PhD, Otsuka, Princeton, NJ, United States; Sven L Klijn, MSc, Bristol Myers Squibb, Princeton, NJ, United States; Howard Thom, PhD, MSc, University of Bristol, Bristol, United Kingdom

SESSION:The complexity, volume, and confidential nature of data in HEOR necessitate the use of innovative tools to synthesize evidence and provide actionable insights. Generative AI solutions, powered by Retrieval-Augmented Generation (RAG), present a reliable approach for managing sensitive data within HEOR workflows.This session will convene experts to discuss the potential of Generative AI-based tools for evidence generation while ensuring complete control over data security. The discussion will cover how security is maintained by implementing Generative AI tools entirely within an organization's infrastructure. Additionally, the session will address the RAG-based data pipeline which can reduce hallucinations and increase reliability, as well as the system's capability to handle various file formats, including text, tables, and images (e.g., PDF, DOC, PPT).Attendees will acquire valuable insights into responsibly integrating RAG and Generative AI into HEOR workflows, thereby facilitating more accessible and evidence-driven decision-making. OVERVIEW: (5 min) The session will be open by Nicola Waddell setting the stage by outlining the potential to revolutionise the generation of reliable, robust evidence, with confidence, using generative AI (5 min) Vakaramoko Diaby will provide an introduction to RAG, elaborating on how this approach uniquely combines generative AI models with external subject-specific data to make it a transformative tool for evidence generation (5 min) Sven Klijn will explore the implications of RAG, discussing how and why it is expected to shape the future of generative AI tools for HEOR use. He will also discuss ongoing developments and innovations that RAG can drive (10 min) Prof. Howard Thom will provide insights from the HTA perspective, focusing on the broader policy implications of integrating innovative tools like RAG into HEOR workflowsAudience Q&A and Panel Discussion (5 min):Attendees will have the opportunity to engage with the panellists, posing questions. Sponsored by Pharmacoevidence Pvt.

Code

103

Topic

Methodological & Statistical Research

Your browser is out-of-date

ISPOR recommends that you update your browser for more security, speed and the best experience on ispor.org. Update my browser now

×