To test and compare the psychometric properties consisting of practicality, reliability, validity and responsiveness of four indirect utility measures including EQ-5D, HUI-2, HUI-3 and SF-6D. The subjects were 216 patients with all types of ischemic heart disease treated at the outpatient clinic of King Chulalongkorn Memorial Hospital. The patients were evaluated health status for two times using four questionnaires including EQ-5D, Health Utilities Index (HUI) which used to calculate the HUI-2 and HUI-3 utility scores, SF-12v2 which used to compute the SF-6D score, and the MacNew Heart disease questionnaire. And all patients were determined the cardiovascular functional class using the Specific Activity Scale. The second evaluation was performed at the next doctor visit of each patient and they were asked for the change of health status by self-rated health question. Regarding the practicality, firstly, we found that the EQ-5D took the lowest time to complete. Secondly, the incomplete rate was found in the HUI-2 and HUI-3 (< 5%). Finally, it was found that the EQ-5D was the only one measure which had the ceiling effect more than the acceptable value (33.3%). Regarding the test-retest reliability, the EQ-5D, HUI-2, and HUI-3 had the high level of ICC (0.798–0.896), while the SF-6D had moderate level (0.709). As for the convergent validity compared with SF-12 and MacNew summary scores, the SF-6D had the highest correlation coefficient (Spearman’s rho > 0.5), while the others were moderate to high correlations. Concerning the known-groups validity, the four measures could discriminate the patients having the different functional class (Kruskal-Wallis test, p < 0.001). Finally, it was found that the SF-6D had the lowest responsiveness ability assessed by the effect size statistics (effect size < 0.2). In conclusion, the EQ-5D had the most suitable psychometric properties. Because it had a good reliability and validity, and it had acceptable responsiveness. However, this study has several limitations for assessing responsiveness. Thus, it needs more evidence to confirm this property.
