Published in:
01-04-2016 | Original Article - Validation Studies
Mapping health assessment questionnaire disability index (HAQ-DI) score, pain visual analog scale (VAS), and disease activity score in 28 joints (DAS28) onto the EuroQol-5D (EQ-5D) utility score with the KORean Observational study Network for Arthritis (KORONA) registry data
Authors:
Hye-Lin Kim, Dam Kim, Eun Jin Jang, Min-Young Lee, Hyun Jin Song, Sun-Young Park, Soo-Kyung Cho, Yoon-Kyoung Sung, Chan-Bum Choi, Soyoung Won, So-Young Bang, Hoon-Suk Cha, Jung-Yoon Choe, Won Tae Chung, Seung-Jae Hong, Jae-Bum Jun, Jinseok Kim, Seong-Kyu Kim, Tae-Hwan Kim, Tae-Jong Kim, Eunmi Koh, Hwajeong Lee, Hye-Soon Lee, Jisoo Lee, Shin-Seok Lee, Sung Won Lee, Sung-Hoon Park, Seung-Cheol Shim, Dae-Hyun Yoo, Bo Young Yoon, Sang-Cheol Bae, Eui-Kyung Lee
Published in:
Rheumatology International
|
Issue 4/2016
Login to get access
Abstract
The aim of this study was to estimate the mapping model for EuroQol-5D (EQ-5D) utility values using the health assessment questionnaire disability index (HAQ-DI), pain visual analog scale (VAS), and disease activity score in 28 joints (DAS28) in a large, nationwide cohort of rheumatoid arthritis (RA) patients in Korea. The KORean Observational study Network for Arthritis (KORONA) registry data on 3557 patients with RA were used. Data were randomly divided into a modeling set (80 % of the data) and a validation set (20 % of the data). The ordinary least squares (OLS), Tobit, and two-part model methods were employed to construct a model to map to the EQ-5D index. Using a combination of HAQ-DI, pain VAS, and DAS28, four model versions were examined. To evaluate the predictive accuracy of the models, the root-mean-square error (RMSE) and mean absolute error (MAE) were calculated using the validation dataset. A model that included HAQ-DI, pain VAS, and DAS28 produced the highest adjusted R
2 as well as the lowest Akaike information criterion, RMSE, and MAE, regardless of the statistical methods used in modeling set. The mapping equation of the OLS method is given as EQ-5D = 0.95−0.21 × HAQ-DI−0.24 × pain VAS/100–0.01 × DAS28 (adjusted R
2 = 57.6 %, RMSE = 0.1654 and MAE = 0.1222). Also in the validation set, the RMSE and MAE were shown to be the smallest. The model with HAQ-DI, pain VAS, and DAS28 showed the best performance, and this mapping model enabled the estimation of an EQ-5D value for RA patients in whom utility values have not been measured.