A mapping study in mainland China: predicting EQ-5D-5L utility scores from the psoriasis disability index

Abstract

Aims: This study aims to develop a mapping algorithm that can be used to predict 5-level EQ-5D (EQ-5D-5L) utility scores from the Psoriasis Disability Index (PDI) for further use in cost-utility analysis among psoriasis patients.Materials and Methods: A total of 321 participants with psoriasis was recruited in a dermatology clinic from a tertiary hospital in China. The EQ-5D-5L utility scores were scored using the Chinese-specific value set. Three model specifications were used and each one was estimated using three different statistical techniques, including the ordinary least squares (OLS), the Tobit regression and the generalized linear model (GLM). The goodness-of-fit of mapping performance was estimated by using two common indicators: the mean absolute error (MAE) and the mean square error (MSE).Results: A final validated sample of 321 psoriasis patients was analyzed in the study. The patients had a mean ± standard deviation (SD) age of 39.7 ± 12.9 years, and the EQ-5D-5L state utility score is 0.899 ± 0.106. Among all the models considered, the OLS linear model based on the four PDI item scores was selected as the best model due to higher accuracy (MAE = 0.1334 and 0.1557 for the estimation and validation sample separately).Limitations: The main limitation of this study is that the survey site was limited to one hospital and not more broadly representing the whole of China. Additionally, there are some missing data in the full sample and an absence of an external dataset to validate the algorithms.Conclusions: The mapping algorithms reported in this paper can be used to predict the EQ-5D-5L utilities from the PDI; however, similar to other mapping studies in the literature it can only serve as the second-best approach to elicit utilities from disease-specific instruments.