Skip to main content

Table 5 Error measurements for predicting utility values based on OHS and OKS questionnaires using the different models. Validation sample

From: Mapping analysis to predict EQ-5D-5 L utility values based on the Oxford Hip Score (OHS) and Oxford Knee Score (OKS) questionnaires in the Spanish population suffering from lower limb osteoarthritis

   MAE MSE SEa ICC
Model Dependent Variable     
Hip osteoarthritis, OHS,n = 301
OLS Utility 0.1343 (0.1215–0.1471) 0.0307 (0.0248–0.0365) 0.0206 (0.0200–0.0213) 0.817 (0.775–0.851)
Tobit Utility 0.1265 (0.1143–0.1387) 0.0275 (0.0220–0.0330) 0.0246 (0.0238–0.0254) 0.833 (0.794–0.864)
GLM Utility 0.1103 (0.0993–0.1214) 0.0216 (0.0167–0.0264) 0.0212 (0.0197–0.0227) 0.855 (0.821–0.882)
Beta reg Utility 0.1229 (0.1102–0.1335) 0.0274 (0.0211–0.0338) 0.0067 (0.0065–0.0070) 0.850 (0.815–0.878)
Knee Osteoarthritis, OKS,n = 316
OLS Utility 0.1278 (0.1159–0.1398) 0.0279 (0.0228–0.0331) 0.0205 (0.0199–0.0211) 0.788 (0.743–0.826)
Tobit Utility 0.1236 (0.1117–0.1355) 0.0268 (0.0216–0.0320) 0.0219 (0.0213–0.0225) 0.791 (0.746–0.829)
GLM Utility 0.1127 (0.1014–0.1239) 0.0230 (0.0181–0.0277) 0.0204 (0.0192–0.0215) 0.824 (0.785–0.856)
Beta reg Utility 0.1141 (0.1031–0.1251) 0.0229 (0.0186–0.0272) 0.0063 (0.0062–0.0064) 0.832 (0.795–0.863)
  1. MAE Mean absolute error, MSE Mean squared error
  2. SEa Standard error reported are the mean values for the original predictions (disutility for GLM and transformed utility (0–1) for beta reg)
  3. CI 95% in parentheses
  4. ICC Intraclass correlation coefficient. Observed-predicted values (absolute agreement)