Are Economic Models in R That Shiny? Making Cost-Effectiveness Models Shinier by Integrating Artificial Intelligence Applications
Author(s)
Devian Parra-Padilla, MSc1, R Lakshmi, MSc2, Akash Yadav, MSc2, Shilpi Swami, MSc1, Tushar Srivastava, MSc1.
1ConnectHEOR, London, United Kingdom, 2ConnectHEOR, Delhi, India.
1ConnectHEOR, London, United Kingdom, 2ConnectHEOR, Delhi, India.
OBJECTIVES: To showcase the integration of generative artificial intelligence (GenAI) into cost-effectiveness modelling in R and R-Shiny aimed at enhancing its development, usability, and dissemination across the pharmaceutical product evidence generation life cycle.
METHODS: An R-based modular cost-effectiveness modelling framework was developed to assess the economic value of a novel hypothetical screening strategy. Both health and economic outcomes were obtained from a healthcare system perspective. The back-end structure in R contained modules for a decision model engine, sensitivity analyses, reactive model inputs, and user-defined settings. These modules were wrapped in a Shiny interface to enable real-time interaction with dynamic graphs of results, modification of inputs, structural assumption testing, and automated output generation. A large language model (LLM) was integrated into the framework via an application programming interface (API). AI-powered report generation was also explored.
RESULTS: The model consisted of 11 distinct modules along with an engine and integration code in the back-end architecture. The integration of R, R Shiny, and GenAI enabled the development of a dynamic and user-friendly modelling tool. The LLM-enhanced application improved the user experience by facilitating the interpretation of the economic modelling framework and results while offering on-demand assistance within the same environment. Leveraging R as an open-source software environment fostered transparency, reproducibility, and accessibility, while addressing limitations of traditional economic modelling tools (such as MS Excel), less compatible with GenAI integration. Further integration is needed to fully automate R Markdown-implemented reporting with LLMs.
CONCLUSIONS: Making Shiny “Shinier” with GenAI opens new opportunities for health economic modelling and can be a relevant tool to enhance the dissemination of economic evidence by supporting policymakers in the decision-making process and ultimately improving patient outcomes.
METHODS: An R-based modular cost-effectiveness modelling framework was developed to assess the economic value of a novel hypothetical screening strategy. Both health and economic outcomes were obtained from a healthcare system perspective. The back-end structure in R contained modules for a decision model engine, sensitivity analyses, reactive model inputs, and user-defined settings. These modules were wrapped in a Shiny interface to enable real-time interaction with dynamic graphs of results, modification of inputs, structural assumption testing, and automated output generation. A large language model (LLM) was integrated into the framework via an application programming interface (API). AI-powered report generation was also explored.
RESULTS: The model consisted of 11 distinct modules along with an engine and integration code in the back-end architecture. The integration of R, R Shiny, and GenAI enabled the development of a dynamic and user-friendly modelling tool. The LLM-enhanced application improved the user experience by facilitating the interpretation of the economic modelling framework and results while offering on-demand assistance within the same environment. Leveraging R as an open-source software environment fostered transparency, reproducibility, and accessibility, while addressing limitations of traditional economic modelling tools (such as MS Excel), less compatible with GenAI integration. Further integration is needed to fully automate R Markdown-implemented reporting with LLMs.
CONCLUSIONS: Making Shiny “Shinier” with GenAI opens new opportunities for health economic modelling and can be a relevant tool to enhance the dissemination of economic evidence by supporting policymakers in the decision-making process and ultimately improving patient outcomes.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
MSR33
Topic
Methodological & Statistical Research
Topic Subcategory
Artificial Intelligence, Machine Learning, Predictive Analytics
Disease
Infectious Disease (non-vaccine)