Abstract / Summary
Abstract Purpose. This study aimed to map health-related quality of life (HRQoL) score into utility score which can be used in cost-effectiveness analysis in Korean gout patients. The research team used HAQ-DI and SF-36v2 to predict EQ-5D-5L utility scores for gout patients. Methods. Adults aged 19–75 with gout from 13 Korean hospitals completed the HAQ-DI, SF-36v2, and EQ-5D-5L at two visits at least one month apart. Utility scores were derived using the Korean EQ-5D-5L value set. Nine regression methods (OLS, Tobit, CLAD, GLM, two-part, beta-mixture, ALDVMM, multinomial logistic, multivariate probit) were applied, with and without age and sex. Models were validated on second-visit data and assessed by MAE, ICC, and absolute-error proportions. Results. The mean EQ-5D-5L utility scores were 0.9114 in first and 0.9281 in second visit. The two-part model was recommended to predict utility scores from HAQ-DI and SF-36v2 (HAQ-DI: RMSE = 0.0863, MAE = 0.0745, AIC = 357.1480/-270.7789, BIC = 364.6212/-262.0409, AE > 0.05(%) = 87.2549, AE > 0.1(%) = 15.6863, ICC = 0.5235) / SF-36v2: RMSE = 0.0877, MAE = 0.0668, AIC = 331.5214/-261.4394, BIC = 342.6327/-249.9386, AE > 0.05(%) = 46.5347, AE > 0.1(%) = 15.8416, ICC = 0.5389 for the internal validation set). The model mapping from HAQ-DI included only HAQ-DI score as the predicted variable, and the model mapping from SF-36v2 included two summary scores (PCS, MCS) as the predicted variables showing the best performance. Conclusion. This study offers validated mapping algorithms for mapping HAQ-DI and SF-36v2 scores into EQ-5D-5L utility scores in Korean patients with gout, supporting evidence generation for clinical and healthcare decision-making.