OBJECTIVES: In HEOR studies, effective collaboration among a multidisciplinary team is essential. Stakeholders from diverse fields including modeling, epidemiology, statistics, clinical practice, pharmacology, and commercial strategy require that HEOR study results be communicated in a manner tailored to their expertise. Complex metrics such as cost per quality-adjusted life years (QALY) or incremental cost-effectiveness ratio (ICER) may not be easily understood by all, especially clinicians unfamiliar with HEOR terms, necessitating summaries in plain English. This study explores the use of Generative-Artificial Intelligence (Gen-AI) to adapt the dissemination of health economic model results to the understanding levels of different target audiences.

METHODS: A proof-of-concept exercise was conducted using ChatGPT 4.0, a language-based Gen-AI implemented in Python. Designed to understand and generate human language, this model was ideal for the task. Multiple virtual stakeholders were defined, each with specific subject knowledge attributes. The Gen-AI was tasked with interpreting results and framing responses suitable for the expertise of various stakeholders (such as modelers, clinicians, providers or payers, layman language, etc.) A human-in-the-loop approach ensured accuracy and relevance of the context.

RESULTS: The Gen-AI effectively interpreted and contextualized results, making them more accessible depending on the stakeholder. It adeptly translated complex HE metrics such as cost per QALY, ICER, and net monetary benefit (NMB) into straightforward terms linked to relevant disease contexts, enhancing understanding for clinicians and non-technical stakeholders.

CONCLUSIONS: This study underscores the potential of Gen-AI in interpreting and communicating the results of HE models tailored to diverse stakeholders. Gen-AI shows promise in bridging communication gaps within HEOR. However, further research is necessary to refine these approaches and fully harness AI's capabilities, ensuring effective and precise information dissemination across varied audiences.