Table 4.
Comparison of the performance of Models 1 and 2
| Model 1a | Model 2b | ||
|---|---|---|---|
| Derivation set |
|
|
|
| Adjusted R2 |
|
0.511 |
0.516 |
| RMSE (% RMSE) |
|
0.095 (8.1) |
0.095 (8.1) |
| MAE (SD) |
|
0.069 (0.065) |
0.069 (0.066) |
| MAE > 0.05(%) |
|
48.7 |
50.1 |
| MAE > 0.1(%) |
|
22.4 |
23.1 |
| EQ-5D index |
Actual |
Predicted |
Predicted |
| Mean (SD) |
0.824 (0.137) |
0.823 (0.098) |
0.824 (0.098) |
| Validation set |
|
|
|
| RMSE (% RMSE) |
|
0.083 (7.1) |
0.085 (7.2) |
| MAE (SD) |
|
0.066 (0.052) |
0.066 (0.053) |
| MAE > 0.05(%) |
|
53.7 |
49.6 |
| MAE > 0.1(%) |
|
23.6 |
24.4 |
| EQ-5D index |
Actual |
Predict |
Predict |
| Mean (SD) | 0.871 (0.113) | 0.873 (0.083) | 0.872 (0.085) |
aModel 1 included all functioning and symptom scales and items of the EORTC QLQ-C30 as explanatory variables.
bModel 2 applied backward elimination to Model 1.