Table 4.
Modelling results for the German EQ-5D-Y value set
| Independent variables of the model | Latent scalea | Rescaledb | ||
|---|---|---|---|---|
| Coefficient | SD | Relative attribute importance (%) | Value set | |
| MO2 | − 0.1778** (0.0767) | 0.1650 (0.1233) | 9.2 | − 0.0242 |
| MO3 | − 0.8627*** (0.1236) | 1.0468*** (0.0936) | − 0.1175 | |
| SC2 | − 0.1401** (0.0566) | 0.4098*** (0.1359) | 11.3 | − 0.0191 |
| SC3 | − 1.0652*** (0.0849) | 0.5188*** (0.1169) | − 0.1450 | |
| UA2 | − 0.6145*** (0.0548) | 0.1687 (0.2060) | 15.5 | − 0.0837 |
| UA3 | − 1.4636*** (0.0845) | 0.5726*** (0.0919) | − 0.1993 | |
| PD2 | − 0.9820*** (0.0594) | 0.0632 (0.0881) | 32.7 | − 0.1337 |
| PD3 | − 3.0772*** (0.1323) | 1.4831*** (0.0976) | − 0.4190 | |
| AD2 | − 0.9213*** (0.0581) | 0.2160 (0.1664) | 31.3 | − 0.1254 |
| AD3 | − 2.9521*** (0.1220) | 1.6490*** (0.0949) | − 0.4019 | |
| Log-likelihood | − 6094 | |||
| Observations | 30,900 | |||
| Sample size | 1030 | |||
Numbers in parentheses represent standard errors
AD feeling worried/sad/unhappy, cTTO composite time trade-off, DCE discrete choice experiment, MO mobility, PD having pain/discomfort, SC looking after myself, SD standard deviation, UA doing usual activities
***p < 0.01, **p < 0.05, *p < 0.1
aBased on a mixed logit model, with all parameters modelled as random and normally distributed, using 5000 Halton draws. Coefficients indicate the decrement from level 1 to the respective level
bRescaled using a linear mapping model between the DCE results and the adjusted mean values from the cTTO task