Comparison of the measurement properties of the EQ-5D-5L and SF-6Dv2 among overweight and obesity populations in China

Table 4 Sensitivity of EQ-5D-5L and SF-6Dv2 to detect differences in different self-reported health status groups (N = 1,000)

Measurement	Categorisation of different self-reported health status groups	N	Utility value (Mean [SD])	t-test		RE^b	ROC curve
Measurement		N	Utility value (Mean [SD])	t-statistic	p-value	RE^b	AUC	95% CI
EQ-5D-5L	excellent	79	0.986 (0.036)	-6.546	< 0.001	1.000	0.815*	(0.784, 0.846)
	good, fair or bad	921	0.839 (0.199)
SF-6Dv2	excellent	79	0.915 (0.101)	-10.758	< 0.001	2.701	0.879*	(0.841, 0.917)
	good, fair or bad	921	0.719 (0.159)
EQ-5D-5L	excellent or good	393	0.952 (0.080)	-14.614	< 0.001	1.000	0.801*	(0.774, 0.827)
	fair or bad	607	0.785 (0.218)
SF-6Dv2	excellent or good	393	0.847 (0.108)	-21.046	< 0.001	2.074	0.862*	(0.839, 0.884)
	fair or bad	607	0.661 (0.152)
EQ-5D-5L	excellent, good or fair	833	0.896 (0.142)	19.478	< 0.001	1.000	0.862*	(0.831, 0.893)
	bad	167	0.622 (0.256)
SF-6Dv2	excellent, good or fair	833	0.773 (0.128)	19.834	< 0.001	1.037	0.881*	(0.853, 0.909)
	bad	167	0.540 (0.183)

Note: *p < 0.001. For the ROC curve, p < 0.001 indicates that AUC is statistically significantly greater than 0.5 and that measure has discriminatory power
a RE of SF-6Dv2 is presented, and reference is EQ-5D-5L, of which RE is 1.000
Abbr: AUC Area under the ROC curve, 95% CI 95% confidence interval, RE Relative efficiency, ROC Receiver operating characteristic, SD Standard deviation

ISSN: 1477-7525