Validation of a Flexible Partitioned Survival and 3-State Markov Model for Cost-Effectiveness Analyses in Oncology
Author(s)
Iwona Zerda, Msc1, Tomasz Fundament, Msc2, Filip Chelmikowski, Msc2, Emilie Clay, Msc, PhD3, Samuel Aballea, MSc, PhD4, Mondher Toumi, MSc, PhD, MD1, Michal Pochopien, MSc, PhD2.
1Aix-Marseille University, Marseille, France, 2Clever-Access, Kraków, Poland, 3Clever-Access, Paris, France, 4Innovintel, Rotterdam, Netherlands.
1Aix-Marseille University, Marseille, France, 2Clever-Access, Kraków, Poland, 3Clever-Access, Paris, France, 4Innovintel, Rotterdam, Netherlands.
OBJECTIVES: Three-state semi-Markov models and partitioned survival models are standard approaches for evaluating new oncological interventions. This study aimed to validate a flexible tool that integrates a partitioned survival model and a 3-state Markov model to simulate the cost-effectiveness of oncology drugs. The tool includes pre-progression, post-progression, and death states, with the flexibility to expand to additional states, such as metastatic disease or second-line treatment. Designed for broad applicability, the tool can be easily populated with inputs relevant to various oncological diseases, enhancing the efficiency of the modelling process.
METHODS: Validation was conducted using data from five recent NICE appraisals of oncology drugs employing the 3-state model structure or its extensions. For each appraisal, the tool was populated with the original inputs and run to replicate published cost-effectiveness results. Key outputs, including clinical and cost outcomes and cost-effectiveness estimates, were compared to the published results to evaluate the accuracy and reliability of the tool.
RESULTS: The reproduced results showed close alignment with the original findings from all five appraisals, confirming the tool’s reliability and robustness. Cost-effectiveness outputs, including ICERs, were consistent with the published appraisals, with minor deviations of < 10%, likely attributed to reporting gaps in the original data.
CONCLUSIONS: This flexible and validated tool demonstrates strong alignment with NICE appraisal results, supporting its accuracy and robustness. Its adaptability allows for rapid input population and expansion to additional health states, significantly reducing development time while maintaining methodological rigor. The tool offers substantial value to researchers and decision-makers in health economics, enabling consistent, accurate, and efficient evaluation of oncological interventions.
METHODS: Validation was conducted using data from five recent NICE appraisals of oncology drugs employing the 3-state model structure or its extensions. For each appraisal, the tool was populated with the original inputs and run to replicate published cost-effectiveness results. Key outputs, including clinical and cost outcomes and cost-effectiveness estimates, were compared to the published results to evaluate the accuracy and reliability of the tool.
RESULTS: The reproduced results showed close alignment with the original findings from all five appraisals, confirming the tool’s reliability and robustness. Cost-effectiveness outputs, including ICERs, were consistent with the published appraisals, with minor deviations of < 10%, likely attributed to reporting gaps in the original data.
CONCLUSIONS: This flexible and validated tool demonstrates strong alignment with NICE appraisal results, supporting its accuracy and robustness. Its adaptability allows for rapid input population and expansion to additional health states, significantly reducing development time while maintaining methodological rigor. The tool offers substantial value to researchers and decision-makers in health economics, enabling consistent, accurate, and efficient evaluation of oncological interventions.
Conference/Value in Health Info
2025-05, ISPOR 2025, Montréal, Quebec, CA
Value in Health, Volume 28, Issue S1
Code
EE358
Topic
Economic Evaluation
Disease
SDC: Oncology