1143P - A comparison of general population and patient utility values for advanced melanoma

Date 01 October 2012
Event ESMO Congress 2012
Session Poster presentation III
Topics Bioethics, Legal, and Economic Issues
Melanoma and other Skin Tumours
Presenter Anthony Batty
Authors A. Batty1, B. Winn1, L. Pericleous2, D. Rowen3, D. Lee4, T. Nikoglou2
  • 1Health Economics, BresMed, S1 2DW - Sheffield/UK
  • 2Health Economics And Outcomes, Bristol Myers Squibb, UB8 1DH - Uxbridge/UK
  • 3School Of Health And Related Research, University of Sheffield, S1 4DA - Sheffield/UK
  • 4BresMed, S1 2DW - Sheffield/UK

Abstract

Background

Health-related quality of life (HRQL) is a fundamental part of health technology assessment. Utility values are vital because they can be used as preference weights and allow the calculation of HRQL benefits. The objective of this study was to compare utilities calculated for patients with advanced melanoma in the Phase III clinical trial for ipilimumab (MDX010-20) using a generic and a condition-specific preference-based measure to utilities produced by vignettes for advanced melanoma. A secondary objective was to determine how different analyses might be used most appropriately within cost-effectiveness modelling.

Methods

The trial utilities were generated using the condition-specific EORTC-8D (1,190 observations) and generic SF-6D (1,157 observations) preference-based measures. Progression-status and time-to-death analyses were conducted on the patient-level data and the predictive abilities were compared. Patient-level results were compared to the utilities derived for progression status using vignettes valued by the general population.

Results

On disease progression, vignette-generated utilities showed a greater decrease (0.77 to 0.59) than either the generic SF-6D (0.64 to 0.619) or condition-specific EORTC-8D (0.801 to 0.763). SF-6D and EORTC-8D showed a large decrease in utility in the 180 days prior to death (from 0.826 to 0.628 and from 0.655 to 0.505, respectively). Compared to progression status, time to death showed a lower Root Mean Squared Error and higher R2 when used to predict patient utility.

Conclusion

Practitioners should carefully analyse patient level utility data prior to constructing economic models as standard measures, such as progression status, may not fully capture the patient experience. Similarly, where vignettes are valued to represent health states in an economic model, their applicability to the disease and patient population should be carefully scrutinised. The use of standard progression based cost-effectiveness modelling techniques may not be appropriate for this disease. Consequently, there are implications for the analysis of utility information in future studies and the methods of cost-effectiveness modelling used for cancer treatments.

Disclosure

A. Batty: BresMed were funded by BMS to conduct the analysis presented within this paper. Other than this I have not received any personal funding and do not have any personal interest in BMS or any competitor products.

B. Winn: BresMed were funded by BMS to conduct the analysis presented within this paper. Other than this I have not received any personal funding and do not have any personal interest in BMS or any competitor products.

L. Pericleous: I have worked for BMS. Other than this I have not received any personal funding and do not have any personal interest in BMS or any competitor products.

D. Lee: BresMed were funded by BMS to conduct the analysis presented within this paper. Other than this I have not received any personal funding and do not have any personal interest in BMS or any competitor products.

T. Nikoglou: I work for BMS. Other than this I have not received any personal funding and do not have any personal interest in BMS or any competitor products.

All other authors have declared no conflicts of interest.