Related Experiment Video
Updated: May 20, 2026
09:29
A Standardized Obstacle Course for Assessment of Visual Function in Ultra Low Vision and Artificial Vision
Published on: February 11, 2014
Estimating quality-adjusted life years from patient-reported visual functioning
C Browne1, J Brazier, J Carlton
1Academic Unit of Health Economics, Leeds Institute of Health Sciences, University of Leeds, Leeds LS2 9LJ, UK. c.browne@leeds.ac.uk
Eye (London, England)
|July 7, 2012
Summary
This study developed algorithms to predict health utility values (EQ-5D, SF-6D) from the Visual Functioning Questionnaire (VFQ-25) and clinical vision tests for glaucoma patients.
Area of Science:
- Ophthalmology
- Health Economics
- Psychometrics
Background:
- Glaucoma significantly impacts patients' health-related quality of life (HRQoL).
- Accurate assessment of HRQoL is crucial for understanding glaucoma's burden.
- Vision-specific and clinical measures are key indicators of visual function.
Purpose of the Study:
- To develop and evaluate mapping algorithms for predicting EQ-5D and SF-6D utility values.
- To utilize the 25-item Visual Functioning Questionnaire (VFQ-25) and clinical visual function tests as predictors.
- To establish a link between vision impairment and broader health utility measures.
Main Methods:
- Comparison of Ordinary Least Squares (OLS), Tobit, and censored least absolute deviations models.
- Utilized data from Moorfields Eye Hospital.
- Model performance assessed using Root Mean Square Error (RMSE), R(2), and Mean Absolute Error (MAE).
Main Results:
- Ordinary Least Squares (OLS) demonstrated superior performance among the compared models.
- OLS yielded the lowest RMSE and MAE, indicating higher accuracy.
- OLS achieved the highest R(2) value, signifying the best model fit.
Conclusions:
- Initial algorithms were established to convert VFQ-25 scores to EQ-5D and SF-6D utility values.
- These algorithms provide a method for estimating health utility from vision-specific data.
- Further validation studies are recommended to confirm the reliability of the developed algorithms.
