Mapping the NEI-VFQ-25 to the EQ-5D-Y-3L and CHU-9D among High School Students with Myopia in China
Author(s)
Chang Luo, MS1, Xiaoyan Yang, PhD2, Jing Wang, MS3, Tianqi Hong, PhD4, Mengdi Li, MS3, Jie Zhang, MS5, Wei Qi, MS6, Lihua Li, MS2, Shitong Xie, PhD1, Jing Wu, PhD1.
1School of Pharmaceutical Science and Technology, Faculty of Medicine, Tianjin University, Tianjin, China, 2Tianjin Eye Hospital, Tianjin, China, 3Institue of Optometry and Vision Sicense, Nankai University, Tianjin, China, 4School of Biomedical Engineering, McMaster University, Hamilton, ON, Canada, 5Heping District Education Comprehensive Service Center, Tianjin, China, 6Jeemei Eye hospital, Qingdao, China.
1School of Pharmaceutical Science and Technology, Faculty of Medicine, Tianjin University, Tianjin, China, 2Tianjin Eye Hospital, Tianjin, China, 3Institue of Optometry and Vision Sicense, Nankai University, Tianjin, China, 4School of Biomedical Engineering, McMaster University, Hamilton, ON, Canada, 5Heping District Education Comprehensive Service Center, Tianjin, China, 6Jeemei Eye hospital, Qingdao, China.
Presentation Documents
OBJECTIVES: To develop algorithms to map the National Eye Institute 25-Item Visual Function Questionnaire (NEI-VFQ-25) onto the EQ-5D-Y-3L and CHU-9D among high school students with myopia in China to inform economic evaluation.
METHODS: Respondents recruited from the Chinese high school myopic students completed an online survey and the sample was randomly divided into development (80%) and validation (20%) datasets. Spearman’s correlation analyses were performed to assess conceptual overlap for NEI-VFQ-25 with EQ-5D-Y-3L and CHU-9D, respectively. Six models, including OLS, Tobit, CLAD, GLM, TPM and ALDVMM, and five predictor sets were explored to estimate mapping algorithms using the development dataset. The mean absolute error (MAE), root mean square error (RMSE), and the proportions of absolute error (AE) within the threshold of 0.05 and 0.10 were used to assess the model performance.
RESULTS: A total of 2,198 students with myopia (53.1% male, mean [SD] age 16.7 [0.8] years) were included in this study. The average score (SD) of NEI-VFQ-25 was 91.112 (9.619) and the mean utility value (SD) of EQ-5D-Y-3L and CHU-9D was 0.962(0.070) and 0.851 (0.160), respectively. The Spearman’s correlation coefficients for NEI-VFQ-25 score with EQ-5D-Y-3L and CHU-9D utilities were 0.366 and 0.502, respectively. The best-performing models predicting EQ-5D-Y-3L and CHU-9D utilities both used the NEI-VFQ-25 subscale score from stepwise regression as predictor set in the OLS model (MAE: 0.043 and 0.101 for the EQ-5D-Y-3L and CHU-9D; RMSE: 0.061 and 0.136 for the EQ-5D-Y-3L and CHU-9D; AE>0.05: 27.0% and 63.71% for the EQ-5D-Y-3L and CHU-9D; AE>0.10: 9.4% and 37.3% for the EQ-5D-Y-3L and CHU-9D).
CONCLUSIONS: OLS models with the NEI-VFQ-25 subscale score from stepwise regression can be used to predict both the EQ-5D-Y-3L and CHU-9D utility values among high school students with myopia in China.
METHODS: Respondents recruited from the Chinese high school myopic students completed an online survey and the sample was randomly divided into development (80%) and validation (20%) datasets. Spearman’s correlation analyses were performed to assess conceptual overlap for NEI-VFQ-25 with EQ-5D-Y-3L and CHU-9D, respectively. Six models, including OLS, Tobit, CLAD, GLM, TPM and ALDVMM, and five predictor sets were explored to estimate mapping algorithms using the development dataset. The mean absolute error (MAE), root mean square error (RMSE), and the proportions of absolute error (AE) within the threshold of 0.05 and 0.10 were used to assess the model performance.
RESULTS: A total of 2,198 students with myopia (53.1% male, mean [SD] age 16.7 [0.8] years) were included in this study. The average score (SD) of NEI-VFQ-25 was 91.112 (9.619) and the mean utility value (SD) of EQ-5D-Y-3L and CHU-9D was 0.962(0.070) and 0.851 (0.160), respectively. The Spearman’s correlation coefficients for NEI-VFQ-25 score with EQ-5D-Y-3L and CHU-9D utilities were 0.366 and 0.502, respectively. The best-performing models predicting EQ-5D-Y-3L and CHU-9D utilities both used the NEI-VFQ-25 subscale score from stepwise regression as predictor set in the OLS model (MAE: 0.043 and 0.101 for the EQ-5D-Y-3L and CHU-9D; RMSE: 0.061 and 0.136 for the EQ-5D-Y-3L and CHU-9D; AE>0.05: 27.0% and 63.71% for the EQ-5D-Y-3L and CHU-9D; AE>0.10: 9.4% and 37.3% for the EQ-5D-Y-3L and CHU-9D).
CONCLUSIONS: OLS models with the NEI-VFQ-25 subscale score from stepwise regression can be used to predict both the EQ-5D-Y-3L and CHU-9D utility values among high school students with myopia in China.
Conference/Value in Health Info
2025-05, ISPOR 2025, Montréal, Quebec, CA
Value in Health, Volume 28, Issue S1
Code
PCR227
Topic
Patient-Centered Research
Topic Subcategory
Health State Utilities, Patient-reported Outcomes & Quality of Life Outcomes
Disease
SDC: Sensory System Disorders (Ear, Eye, Dental, Skin)