Health economic models (HEMs) are indispensable tools for assessing the cost-effectiveness of healthcare interventions. These models provide critical insights into the relative costs and benefits of new health technologies compared to existing alternatives, thereby informing health policy decisions, reimbursement strategies, and market access for pharmaceuticals and medical devices. However, as healthcare pathways become increasingly complex, the traditional methods for developing HEMs are being challenged. In this context, there is a growing need for innovative approaches that can not only handle this complexity but also uncover unseen patterns and generate novel solutions that add significant value.

In recent years, advances in artificial intelligence (AI), particularly with large language models (LLMs) such as the Generative Pre-trained Transformer 4 (GPT-4), have shown potential to revolutionize various domains by automating complex cognitive and reasoning tasks. The aim of this session is to discuss how LLMs can be utilised in HEOR reasoning problems, specifically related to conceptualising the natural history of the disease and recommending potential model structures.

The session will cover the following topics:

• Overview of human way of model conceptualisation for de novo models
• AI integration frameworks and approaches (such as human in loop, AI in loop)
• Reasoning algorithms for LLMs to be utilized for reasoning problems such as model conceptualization (e.g., chain of thoughts, tree of thoughts, graph of thoughts)
• Overview of HEM-X^TM – a proprietary tool of ConnectHEOR trained for model conceptualization (using a case study demonstration)
• Further developments and scope of LLMs for reasoning problems in HEOR
  
  

Sponsor: ConnectHEOR