• Nick Kontodimopoulos, Vassilis H Aletras, Dimitris Paliouras, and Dimitris Niakas.
• Faculty of Social Sciences, Hellenic Open University, Bouboulinas 57, Patras, Greece. nkontodi@otenet.gr
• Value Health. 2009 Nov 1; 12 (8): 1151-7.

ObjectivesTo estimate models, via ordinary least squares regression, for predicting Euro Qol 5D (EQ-5D), Short Form 6D (SF-6D), and 15D utilities from scale scores of the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ-C30).MethodsForty-eight gastric cancer patients, split up into equal subgroups by age, sex, and chemotherapy scheme, were interviewed, and the survey included the QLQ-C30, SF-36, EQ-5D, and 15D instruments, along with sociodemographic and clinical data. Model predictive ability and explanatory power were assessed by root mean square error (RMSE) and adjusted R(2) values, respectively. Pearson's r between predicted and reported utility indices was compared. Three random subsamples, half in size the initial sample, were created and used for "external" validation of the modeling equations.ResultsExplanatory power was high, with adjusted R(2) reaching 0.909, 0.833, and 0.611 for 15D, SF-6D, and EQ-5D, respectively. After normalization of RMSE to the range of possible values, the prediction errors were 12.0, 5.4, and 5.6% for EQ-5D, SF-6D, and 15D, respectively. The estimation equations produced a range of utility scores similar to those achievable by the standard scoring algorithms. Predicted and reported indices from the validation samples were comparable thus confirming the previous results.ConclusionsEvidence on the ability of QLQ-C30 scale scores to validly predict 15D and SF-6D utilities, and to a lesser extent, EQ-5D, has been provided. The modeling equations must be tried in future studies with larger and more diverse samples to confirm their appropriateness for estimating quality-adjusted life-year in cancer-patient trials including only the QLQ-C30.

29 keywords...

hide…