Utility Over Multiple Health Indexes in Non-Small Cell Lung Cancer: Survey Estimates for Patient-Centered Generalized Risk-Adjusted Cost-Effectiveness (GRACE)
Author(s)
Rahul Mudumba, MHS, Karen Mulligan, PhD, Jorge J Nieva, MD, Darius Lakdawalla, PhD;
University of Southern California, Los Angeles, CA, USA
University of Southern California, Los Angeles, CA, USA
Presentation Documents
OBJECTIVES: Traditional cost-effectiveness analysis (CEA) assumes risk-neutrality over health, conflicting with conventional economic principles like diminishing returns to health and risk-aversion. Generalized Risk-Adjusted Cost-Effectiveness (GRACE) relaxes this assumption, but effective implementation requires utility estimates that capture risk preferences varying over health levels. This study aims to directly estimate utility functions for metastatic non-small cell lung cancer (mNSCLC) patients using multiple health indexes, providing a broader library of health measurement and utility function choices for GRACE practitioners.
METHODS: We will develop, pilot, and disseminate a survey to estimate utility and risk preferences over health from mNSCLC patients recruited from patient registries and clinical practices. Leveraging prior empirical research on GRACE utility estimation, the survey will include 12 hypothetical health gamble scenarios, split between Visual Analog Scale (VAS) and Time Trade-Off (TTO) indexes, to elicit certainty equivalents (CEs) that identify indifference between risky and certain health outcomes. Based on these CEs, utility functions will be estimated under four parametric models: constant-relative risk aversion (CRRA), hyperbolic absolute relative risk aversion (HARA), and one- and two-parameter expo-power (EP).
RESULTS: Although data collection is ongoing, we anticipate generating patient-derived utility functions that can be used both with VAS and TTO health measures; preliminary results are anticipated by the time of the conference. These functions will provide analysts the estimates necessary to conduct patient-centric and tailored disease-specific GRACE analyses in an array of contexts. Furthermore, utility functions derived from multiple health indexes enable use of GRACE whenever the analyst has health measures in terms of VAS or TTO.
CONCLUSIONS: This study represents one of the first efforts to estimate utility over health directly from patients using multiple health indexes. By enabling analysts to integrate patient preferences into GRACE and simplifying implementation when indexes align, this work advances patient-centered evaluations and supports more accurate, comprehensive, and equitable value assessments.
METHODS: We will develop, pilot, and disseminate a survey to estimate utility and risk preferences over health from mNSCLC patients recruited from patient registries and clinical practices. Leveraging prior empirical research on GRACE utility estimation, the survey will include 12 hypothetical health gamble scenarios, split between Visual Analog Scale (VAS) and Time Trade-Off (TTO) indexes, to elicit certainty equivalents (CEs) that identify indifference between risky and certain health outcomes. Based on these CEs, utility functions will be estimated under four parametric models: constant-relative risk aversion (CRRA), hyperbolic absolute relative risk aversion (HARA), and one- and two-parameter expo-power (EP).
RESULTS: Although data collection is ongoing, we anticipate generating patient-derived utility functions that can be used both with VAS and TTO health measures; preliminary results are anticipated by the time of the conference. These functions will provide analysts the estimates necessary to conduct patient-centric and tailored disease-specific GRACE analyses in an array of contexts. Furthermore, utility functions derived from multiple health indexes enable use of GRACE whenever the analyst has health measures in terms of VAS or TTO.
CONCLUSIONS: This study represents one of the first efforts to estimate utility over health directly from patients using multiple health indexes. By enabling analysts to integrate patient preferences into GRACE and simplifying implementation when indexes align, this work advances patient-centered evaluations and supports 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
PCR142
Topic
Patient-Centered Research
Topic Subcategory
Health State Utilities, Patient-reported Outcomes & Quality of Life Outcomes
Disease
SDC: Oncology