Skip to main content

Table 4 Comparison of the 5 models based on the combined set

From: Results from an exploratory study to test the performance of EQ-5D-3L valuation subsets based on orthogonal designs, and an investigation into some modeling and transformation alternatives for the utility function

  Baseline model N1 N2 N3 C3Sq
Constant 0.8685 0.9830 0.8865 0.9279 0.8921
MO2 −0.1350 −0.1437 −0.1084 −0.1442 −0.1395
MO3 −0.3874 −0.4061 −0.3725 −0.3749 −0.4325
SC2 −0.0975 −0.1061 −0.0766 −0.0876 −0.0911
SC3 −0.1614 −0.1802 −0.1581 −0.1394 −0.2023
UA2 −0.0764 −0.0850 −0.0613 −0.0475 −0.0674
UA3 −0.0988 −0.1176 −0.1070 −0.0673 −0.1322
PD2 −0.0495 −0.0650 −0.0289 −0.0587 −0.0523
PD3 −0.1390 −0.1510 −0.1362 −0.1265 −0.1807
AD2 −0.0410 −0.0426 −0.0202 −0.0494 −0.0420
AD3 −0.0849 −0.1108 −0.0816 −0.0720 −0.1282
N1   −0.0831    
N2    −0.0709   
N3     −0.1119  
C3Sq      0.0100
Adj R Sq 0.5550 0.5597 0.5588 0.5647 0.5597
Correl Model vs Observed 0.9816 0.9847 0.9840 0.9881 0.9867
% MAE Model vs Observed 3.0% 3.1% 2.7% 2.4% 2.7%
% MAE Model vs Holdouts 2.8% 2.8% 3.5% 1.5% 2.6%
% Residuals >0.05 11% 13% 21% 11% 13%