Unstable preferences: a shift in valuation or an effect of the elicitation procedure?

OBJECTIVE: Many studies suggest that impaired health states are valued more positively when experienced than when hypothetical. This study investigated to what extent this discrepancy occurs and examined four possible explanations: non-corresponding description of the hypothetical health state, new understanding due to experience with the health state, valuation shift due to a new status quo, and instability of preference. PATIENTS AND METHODS: Fifty-five breast cancer patients evaluated their actually experienced health state, a radiotherapy scenario, and a chemotherapy control scenario before, during, and after postoperative radiotherapy. Utilities were elicited by means of a visual analog scale (VAS), a chained time tradeoff (TTO), and a chained standard gamble (SG).
RESULTS: The discrepancy was found for all methods and was statistically significant for the TTO (predicted utilities: 0.89, actual utilities: 0.92, p < or = 0.05). During radiotherapy, significant differences (p < or = 0.01) were found between the utilities for the radiotherapy scenario and the actual health state by means of the VAS and the SG, suggesting non-corresponding description as an explanation. The utilities of the radiotherapy scenario and the chemotherapy control scenario remained stable over time, and thus new understanding, valuation shift, and instability could be ruled out as explanations.
CONCLUSION: Utilities obtained through hypothetical scenarios may not be valid predictors of the value judgments of actually experienced health states. The discrepancy in this study seems to have been due to differences between the situations in question (non-corresponding descriptions).