P F M Stalmeier1, A L Verheijen. 1. Department for Health Evidence, Radboud University Medical Centre, HEV 133, P.O. Box 9101, 6500 HB, Nijmegen, The Netherlands, P.Stalmeier@ebh.umcn.nl.
Abstract
BACKGROUND: The time trade off (TTO) method is not sensitive to maximal endurable time preferences, as preference reversals occur. The standard gamble (SG) method has not been tested regarding its sensitivity to maximal endurable time preferences. OBJECTIVE: This study investigates whether preference reversals occur for the SG method as well. METHODS: Fifty-nine respondents stated for several migraine health states their preference for living 10 or 20 years in that state. A migraine state was selected for which a respondent preferred 10-20 years, a maximal endurable time preference. Two probability equivalent gambles were obtained for the migraine states lasting 10 and 20 years, respectively. Preference reversals occurred when the gamble, equivalent to the longer duration, was preferred to the gamble equivalent to the shorter duration. RESULTS: Out of 59 respondents, 48 had maximal endurable time preferences. Of these 48 respondents, 34 (71%) showed a preference reversal. This percentage differed significantly from chance, that is 50% (P = 0.004), indicating that preference reversals occurred reliably. CONCLUSION: The observed reversal rate for the standard gamble is similar to rates observed previously with the TTO method. Utility measurement of poor health states is problematic, both with the TTO and standard gamble methods.
BACKGROUND: The time trade off (TTO) method is not sensitive to maximal endurable time preferences, as preference reversals occur. The standard gamble (SG) method has not been tested regarding its sensitivity to maximal endurable time preferences. OBJECTIVE: This study investigates whether preference reversals occur for the SG method as well. METHODS: Fifty-nine respondents stated for several migraine health states their preference for living 10 or 20 years in that state. A migraine state was selected for which a respondent preferred 10-20 years, a maximal endurable time preference. Two probability equivalent gambles were obtained for the migraine states lasting 10 and 20 years, respectively. Preference reversals occurred when the gamble, equivalent to the longer duration, was preferred to the gamble equivalent to the shorter duration. RESULTS: Out of 59 respondents, 48 had maximal endurable time preferences. Of these 48 respondents, 34 (71%) showed a preference reversal. This percentage differed significantly from chance, that is 50% (P = 0.004), indicating that preference reversals occurred reliably. CONCLUSION: The observed reversal rate for the standard gamble is similar to rates observed previously with the TTO method. Utility measurement of poor health states is problematic, both with the TTO and standard gamble methods.