Intensity of chemotherapy-induced emesis and overall survival as determinants of a global utility score

Abstract

Cost-utility analysis is appropriate for supportive care interventions, since both quality of life and duration of survival are taken into account. However, utility scores, which reflect quality of life, have not been objectively defined for parameters such as emesis. The Standard Gamble technique identifies utility by determining patient preference for retaining a fixed intermediate health state or taking a gamble with perfect health and death as possible outcomes (the odds of the two gamble outcomes can be varied). If death is not a reasonable negative anchor (for example, with transient symptoms of treatment), then a surrogate negative anchor that is itself defined according to a perfect health/death gamble may be used. We explored use of the Standard Gamble to define a reasonable surrogate negative anchor for evaluation of the impact of emesis. Ten normal volunteers performed a Standard Gamble exercise in which the fixed health state was a year of chemotherapy followed by survival in good health and the gamble had possible outcomes of perfect health or immediate death. The amount of nausea/vomiting during chemotherapy (continuous or none) and the duration of overall survival (6 years or 2 years) were varied to produce four unique health states. All subjects completed the exercise and demonstrated construct validity. Mean utility scores ranged from 0.89 for 6-year survival/no emesis to 0.46 for 2-year survival/continuous emesis. Presence of emesis significantly affected utility (0.565 vs 0.88), while shorter survival had a lesser effect (0.67 vs 0.78). The Standard Gamble technique can be used to evaluate nausea and vomiting. The fixed health state of 1 year of chemotherapy with continuous nausea and vomiting followed by 1 year of good health defines a utility score of 0.46. This is a reasonable negative surrogate anchor and is being used in a current Cancer and Leukemia Group B (CALGB) study.