Skip to main content

Table 3 Predictive accuracy of the algorithms by the observed UK EQ-5D-3L index score

From: Assessing the external validity of algorithms to estimate EQ-5D-3L from the WOMAC

EQ-5D-3L index n   Barton Xie Wailoo
      WA CEC
<0.5 100 ME (95 % CI) 0.308 (0.267 to 0.349) 0.551 (0.529 to 0.574) 0.256 (0.209 to 0.303) 0.219 (0.151 to 0.288)
MAE 0.325 0.551 0.281 0.324
RMSE 0.372 0.563 0.350 0.410
0.5–0.699 193 ME (95 % CI) −0.111 (−0.130 to −0.091) 0.057 (0.049 to 0.065) −0.144 (−0.169 to −0.119) −0.126 (−0.165 to −0.086)
MAE 0.142 0.063 0.187 0.194
RMSE 0.177 0.079 0.228 0.307
0.7–0.899 645 ME (95 % CI) −0.098 (−0.106 to −0.089) −0.013 (−0.017 to −0.009) −0.094 (−0.106 to −0.081) −0.054 (−0.070 to −0.038)
MAE 0.110 0.038 0.147 0.139
RMSE 0.147 0.048 0.187 0.211
0.9–1.0 130 ME (95 % CI) −0.215 (−0.223 to −0.206) −0.183 (−0.188 to −0.179) −0.096 (−0.113 to −0.079) −0.062 (−0.081 to −0.042)
MAE 0.215 0.183 0.096 0.062
RMSE 0.221 0.185 0.137 0.131
All 1068 ME (95 % CI) −0.076 (−0.087 to −0.066) 0.032 (0.021 to 0.043) −0.070 (−0.082 to −0.058) −0.042 (−0.057 to −0.028)
MAE 0.148 0.108 0.161 0.157
RMSE 0.194 0.191 0.210 0.249
Absolute error > |0.05 observed|, % 78.8 58.2 82.8 75.3
Absolute error > |0.10 observed|, % 64.7 36.4 65.9 53.3
Absolute error > |0.25 observed|, % 29.7 11.3 36.2 24.3
  1. CI confidence interval, ME mean error, MAE mean absolute error, RMSE root mean square error, WA weighted average, CEC conditional on estimated component