A Simple and Practical Guide to Implementing Generalized Risk-Adjusted Cost-Effectiveness (GRACE): A Case Study in Non-Small Cell Lung Cancer
Author(s)
Rahul Mudumba, MHS, William V. Padula, PhD, Darius Lakdawalla, PhD;
University of Southern California, Los Angeles, CA, USA
University of Southern California, Los Angeles, CA, USA
Presentation Documents
OBJECTIVES: Generalized Risk-Adjusted Cost-Effectiveness (GRACE) offers a robust and comprehensive value assessment framework, yet perceived complexity limits its uptake and may lead to shortcuts resulting in flawed execution. This study aims to outline a simplified process for implementing GRACE and demonstrates its application by extending a traditional cost-effectiveness analysis (CEA) of three therapies for advanced non-small cell lung cancer (NSCLC) into a GRACE analysis.
METHODS: A traditional CEA model evaluating alectinib, brigatinib, and lorlatinib in advanced anaplastic lymphoma kinase (ALK)-positive NSCLC was extended into GRACE. Health-related quality of life (HRQoL) values for each modeled health state were converted into risk-adjusted utility values using utility functions estimated in societal and patient-centric risk preference surveys. Next, these risk-adjusted values were introduced into the previous CEA model to enable GRACE analysis. Comparisons were made between traditional CEA and GRACE results, examining the effects of societal versus patient-derived risk preferences and potential shortcuts in implementation. Statistical significance was assessed at the 5% level using Monte Carlo simulations over 10,000 iterations.
RESULTS: GRACE analysis, based on societal utility functions and implementation shortcuts, demonstrated systematic shifts in incremental cost-effectiveness ratios (ICERs) over 10,000 iterations. For alectinib vs. brigatinib, the median ICER decreased from $248,990/quality-adjusted life year (QALY) under the traditional model to $232,568/QALY under GRACE (p < 0.001), representing a 7% reduction. Similarly, the median ICER for lorlatinib vs. brigatinib decreased from $481,635/QALY to $420,038/QALY (p < 0.001), a 13% reduction. Across both comparisons, GRACE adjustments showed consistent effects over iterations. Results incorporating patient-centric preferences will be presented at the conference following completion of an ongoing study.
CONCLUSIONS: GRACE analyses can significantly shift cost-effectiveness conclusions and accommodate both societal and patient-centered risk preferences. This study provides a step-by-step guide to extending traditional CEA models into GRACE, demonstrating its feasibility and flexibility in advancing more accurate, comprehensive, and equitable value assessments.
METHODS: A traditional CEA model evaluating alectinib, brigatinib, and lorlatinib in advanced anaplastic lymphoma kinase (ALK)-positive NSCLC was extended into GRACE. Health-related quality of life (HRQoL) values for each modeled health state were converted into risk-adjusted utility values using utility functions estimated in societal and patient-centric risk preference surveys. Next, these risk-adjusted values were introduced into the previous CEA model to enable GRACE analysis. Comparisons were made between traditional CEA and GRACE results, examining the effects of societal versus patient-derived risk preferences and potential shortcuts in implementation. Statistical significance was assessed at the 5% level using Monte Carlo simulations over 10,000 iterations.
RESULTS: GRACE analysis, based on societal utility functions and implementation shortcuts, demonstrated systematic shifts in incremental cost-effectiveness ratios (ICERs) over 10,000 iterations. For alectinib vs. brigatinib, the median ICER decreased from $248,990/quality-adjusted life year (QALY) under the traditional model to $232,568/QALY under GRACE (p < 0.001), representing a 7% reduction. Similarly, the median ICER for lorlatinib vs. brigatinib decreased from $481,635/QALY to $420,038/QALY (p < 0.001), a 13% reduction. Across both comparisons, GRACE adjustments showed consistent effects over iterations. Results incorporating patient-centric preferences will be presented at the conference following completion of an ongoing study.
CONCLUSIONS: GRACE analyses can significantly shift cost-effectiveness conclusions and accommodate both societal and patient-centered risk preferences. This study provides a step-by-step guide to extending traditional CEA models into GRACE, demonstrating its feasibility and flexibility in advancing more accurate, comprehensive, and equitable value assessments.
Conference/Value in Health Info
2025-05, ISPOR 2025, Montréal, Quebec, CA
Value in Health, Volume 28, Issue S1
Code
EE192
Topic
Economic Evaluation
Disease
SDC: Oncology, SDC: Rare & Orphan Diseases