Table 1 Comparison of valuation method and characteristics of the two EQ-5D-3L value sets for China

Valuation method

Sample size used

1222 respondents

6000 respondents

Sampling area

Beijing, Shenyang, Nanjing, Chengdu, and Guangzhou (Urban area)

Jiangsu, Guangdong, Hebei, Chongqing, and Shaanxi (One rural and one urban area)

Time of data collection

2011.03.11–05.25

2014.07.10–08.25

Sampling method

Quota sampling

A multistage, stratified, clustered random sampling

Number of health states directly valuated

97

43

Number of health states valued by each respondents

13

13

Valuation protocol used

Paris protocol

MVH protocol

Modeling approach

Ordinary least squares; weighted least squares

Ordinary least squares; general least squares; weighted least squares

Choice of final model

An ordinary least square model including 10 dummies with constant and N3

An ordinary least square model including 10 dummies without constant and N3

Characteristics of the two value sets

The range of index scores

[−0.149, 1]

[0.170, 1]

The median of index scores

0.427

0.653

Number of health states worse than dead (%)

6 (2.5%)

0 (0%)

Dimension importance order

MO, PD, SC, AD, UA

SC, MO, AD, UA, PD

Scoring parameter

1−(0.039 + 0.099*MO2 + 0.246*MO3 + 0.105*SC2 + 0.208*SC3 + 0.074*UA2 + 0.193*UA3 + 0.092*PD2 + 0.236*PD3 + 0.086*AD2 + 0.205*AD3 + 0.022*N3)

1−(0.077*MO2 + 0.267*MO3 + 0.044*SC2 + 0.291*SC3 + 0.037*UA2 + 0.054*UA3 + 0.027*PD2 + 0.041*PD3 + 0.036*AD2 + 0.177*AD3)