Mapping ODI onto EQ-5D-5L in Chinese Low Back Pain Patients

Abstract

Mapping can translate utility values from other health-related quality-of-life scales, giving researchers and policymakers more comprehensive information. The primary objective of the study is to develop mapping algorithms that convert scores from the Oswestry Disability Index (ODI) to the 5-level EuroQol-5 Dimension (EQ-5D-5L). Data for this analysis was sourced from 272 patients suffering from low back pain. The development of the mapping algorithms involved the application of three distinct regression methods across four different settings: ordinary least squares regression, beta regression, and multivariate ordered probit regression. To evaluate the internal validity of these algorithms, we adopted a 'hold-out' approach for predictive performance assessment. Furthermore, to discern the most effective model, three goodness-of-fit tests were employed: the mean absolute error (MAE), the root-mean-square error (RMSE), and the Spearman rank correlation coefficients between the predicted and observed utilities. The study successfully developed several models capable of accurately predicting health utilities in the specified context. The best performing models for ODI to EQ-5D-5L mapping were beta regressions. Mapping algorithms developed in this study enable the estimation of utility values from the ODI. The algorithms formulated in this study facilitate the estimation of utility values based on the ODI, providing a valuable empirical foundation for estimating health utilities in scenarios where EQ-5D data is unavailable.