Which approach is better in eliciting health state utilities from breast cancer patients? Evidence from mainland China.

Accurately measuring health state utilities is crucial for health economic evaluation. This study empirically investigated both indirect approaches (including using the latest version of five-level EQ-5D questionnaire, EQ-5D-5L and the popular SF-6D), as well as a direct approach (the time trade-off, TTO) for eliciting utilities in breast cancer patients. A valid sample of 608 breast cancer patients (48% TNM Stages III/IV) in mainland China was analysed. Mean utilities elicited from the TTO, EQ-5D-5L and SF-6D were 0.80, 0.83 and 0.65 respectively. There is poor to fair agreement between direct and indirect approaches on measuring health state utilities with breast cancer patients. The absolute agreement was higher between EQ-5D-5L and SF-6D utilities (intraclass correlation coefficient, ICC = 0.55) than between TTO and EQ-5D-5L/SF-6D utilities (ICCs < 0.3). Regression analyses found that both EQ-5D-5L and SF-6D utilities were responsive to five out of eight breast cancer-specific symptom and functional scales. In sum, criterion and known-group validities of three health state utility measures which were investigated in this research suggest that, the EQ-5D-5L is currently the optimal approach (followed by the SF-6D) to elicit health state utilities from breast cancer patients in mainland China.