Session (CTI)

Separate registration required.

Generative AI (GenAI) is rapidly transforming how health economics and outcomes research (HEOR) is conducted from literature reviews and evidence synthesis to economic modeling and HTA submissions. As the field moves beyond experimentation, professionals face a new challenge: how to responsibly validate, implement, and scale GenAI solutions in real-world HEOR settings.

This intermediate-level course builds upon basic concepts and is designed for HEOR professionals, data scientists, and decision makers seeking to understand not only how GenAI works, but how to implement and evaluate it effectively within regulated and evidence-driven environments. The course provides a practical framework for moving “from prototype to practice,” describing the lifecycle of GenAI implementation—from early sprints and pilot projects to production deployment. Participants will explore both technical and organizational perspectives, including workflow orchestration, modularization, scaling, and change management.

Retrieval-Augmented Generation (RAG) is a cornerstone architecture that integrates external knowledge bases into LLM workflows. Faculty will discuss why RAG is particularly relevant for HEOR, demonstrating how external information (eg, clinical data, published evidence, HTA guidance) can be incorporated in GenAI workflows according to best practice standards and used to improve factual accuracy and traceability. A guided practical session is included so participants become familiar with how to implement a simple RAG pipeline, learning how to chunk data, generate embeddings, and augment prompts for domain-specific use. The course will also provide an extensive overview of Agentic AI, a fast-evolving frontier in AI automation. Participants will examine how autonomous AI “agents” can coordinate multi-step HEOR processes—such as literature updates, model maintenance, or simulated committee reviews—while maintaining control and accountability. A second practical session will demonstrate an agentic workflow in action, showcasing task orchestration, monitoring, and boundary setting. Beyond technical topics, there will be a focus on evaluation and validation of GenAI solutions for HEOR, where participants will learn how to critically assess GenAI tools in terms of reliability, reproducibility, and regulatory alignment. This will also be discussed in the context of potential ethical concerns around the application of AI. Using frameworks such as ELEVATE-GenAI, and referencing NICE and FDA guidance, participants will learn how to ensure that AI-driven outputs meet HEOR’s high standards for quality and transparency. By the end of this course, participants will understand how to bridge the gap between exploratory AI use and operational excellence. They will leave with actionable frameworks and hands-on knowledge to evaluate, implement, and govern GenAI tools that enhance productivity, transparency, and scientific integrity across HEOR activities.

PREREQUISITES: Completion of the “Applied Generative AI for HEOR: Introduction” ISPOR course or familiarity with concepts such as prompt engineering, APIs, and LLM workflows. A basic understanding of Python or other similar scripting languages is recommended to get the most benefit from the guided practical sessions.