Country-Specific Adaptation of Health Economic Models With Generative AI: A Case Study in Alzheimer’s Disease
Author(s)
Jag Chhatwal, PhD1, Sumeyye Samur, PhD2, Ismail Fatih Yildirim, MSc3, Jade Xiao, PhD3, Jamie Elvidge, BA, MSc4, Kusal Lokuge, PhD5, Steve Sharp, MSc4, Rachael Fleurence, MSc, PhD3, Turgay Ayer, PhD6.
1Harvard Medical School / Massachusetts General Hospital, Boston, MA, USA, 2VP, Head of Value & Access, Value Analytics Labs, Boston, MA, USA, 3Value Analytics Labs, Boston, MA, USA, 4National Institute for Health and Care Excellence, Manchester, United Kingdom, 5National Institiute for Health and Care Excellence, Manchester, United Kingdom, 6Georgia Institute of Technology, Atlanta, GA, USA.
1Harvard Medical School / Massachusetts General Hospital, Boston, MA, USA, 2VP, Head of Value & Access, Value Analytics Labs, Boston, MA, USA, 3Value Analytics Labs, Boston, MA, USA, 4National Institute for Health and Care Excellence, Manchester, United Kingdom, 5National Institiute for Health and Care Excellence, Manchester, United Kingdom, 6Georgia Institute of Technology, Atlanta, GA, USA.
OBJECTIVES: Generative AI (GenAI) offers a novel approach to streamlining the adaptation of health economic models for different country contexts. Traditional adaptation processes can be time- and resource-intensive, requiring the identification of local input parameters. This study explores the application of GenAI in adapting a US-based Alzheimer’s cost-effectiveness model to the UK setting.
METHODS: We implemented a stepwise process using GenAI to adapt a US-based Alzheimer’s disease model from the Institute for Clinical and Economic Review (ICER) to the UK. We developed a roadmap for identifying critical input domains, querying appropriate data sources, and translating findings into model-ready formats using AI. Human oversight was used to select context-appropriate parameters, validate AI outputs, and ensure consistency with established health economic modeling practices. Key input domains included: unit costs, health state utilities, and mortality. Publicly available UK-specific data sources were prioritized, including NICE reports, OECD purchasing power parity (PPP) indices, and WHO mortality databases. GenAI, with human-in-the-loop, was used to identify, extract and translate inputs, update model parameters, and run the adapted model.
RESULTS: The AI-supported process enabled efficient, transparent model adaptation. GenAI was partially successful in identifying UK-specific inputs, including health utility values and costs, applying PPP-based cost conversions, and filling gaps in mortality data with global averages. While human-in-the-loop was needed to guide parameter selection, GenAI updated the model and ran the adapted model without human oversight.
CONCLUSIONS: This case study illustrates how GenAI can support country-specific adaptation of health economic models. While the approach shows promise, further research is needed to quantify its benefits and assess its generalizability across different model types, disease areas and jurisdictions. These findings can inform the development of a scalable roadmap to guide AI-supported model adaptation.
METHODS: We implemented a stepwise process using GenAI to adapt a US-based Alzheimer’s disease model from the Institute for Clinical and Economic Review (ICER) to the UK. We developed a roadmap for identifying critical input domains, querying appropriate data sources, and translating findings into model-ready formats using AI. Human oversight was used to select context-appropriate parameters, validate AI outputs, and ensure consistency with established health economic modeling practices. Key input domains included: unit costs, health state utilities, and mortality. Publicly available UK-specific data sources were prioritized, including NICE reports, OECD purchasing power parity (PPP) indices, and WHO mortality databases. GenAI, with human-in-the-loop, was used to identify, extract and translate inputs, update model parameters, and run the adapted model.
RESULTS: The AI-supported process enabled efficient, transparent model adaptation. GenAI was partially successful in identifying UK-specific inputs, including health utility values and costs, applying PPP-based cost conversions, and filling gaps in mortality data with global averages. While human-in-the-loop was needed to guide parameter selection, GenAI updated the model and ran the adapted model without human oversight.
CONCLUSIONS: This case study illustrates how GenAI can support country-specific adaptation of health economic models. While the approach shows promise, further research is needed to quantify its benefits and assess its generalizability across different model types, disease areas and jurisdictions. These findings can inform the development of a scalable roadmap to guide AI-supported model adaptation.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
MSR65
Topic
Economic Evaluation, Health Technology Assessment, Methodological & Statistical Research
Topic Subcategory
Artificial Intelligence, Machine Learning, Predictive Analytics
Disease
Neurological Disorders