Purchase this article with an account.
Y. Alavi, M. Jofre-Bonet, C. Bunce, R. Wormald, A. C. Viswanathan, A. Foster, R. A. Hitchings; Developing an Algorithm to Predict Generic Health-Related Quality of Life Utility Values From Routine Measures of Visual Function in Primary Open Angle Glaucoma. Invest. Ophthalmol. Vis. Sci. 2008;49(13):1159.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To develop an algorithm that can predict health-related quality of life utility values for patients with primary open angle glaucoma (POAG) using combinations of routine measures of binocular visual acuity (VAB), contrast sensitivity (CSB), and visual field.
132 POAG patients and 14 controls were recruited. Participants completed the Time Trade-off (TTO) utility question, 25-item National Eye Institute Visual Function Questionnaire, Daily Living Tasks dependent on Vision, EuroQoL EQ-5D, and the 36 item Short Form (SF-36) instruments during face-to-face interviews. ETDRS logMAR visual acuity, Pelli-Robson contrast sensitivity and Humphrey monocular 24-2 visual field scores were measured on the same day. Integrated (binocular) Visual Field (IVF) scores were derived from the latter perimetry tests. Tobit regression analyses were used to model utility values based on combinations of binocular visual function measures and other controlling variables.
Utility values recorded for 123 cases were significantly lower than those recorded for controls (P < 0.05) and demonstrated good construct validity through significant correlation with both clinical measures of binocular visual function (r = -0.47, IVF; r = -0.48, VAB; r = 0.50, CSB; P <0.0001) and patient-reported measures of vision-specific quality of life (r = 0.54-0.6, P <0.0001). Weaker correlations were observed with generic health measures whose constructs did not target vision-specific quality of life (r = 0.21 - 0.27, P 0.05, SF-36 mental health component summary score). Two final predictive models incorporate terms for visual field and visual acuity, with or without living arrangements, and explain up to 34% of variation in utilities. Contrast sensitivity was not included in either model due to co-linearity between CSB and VAB.
Further validation of the above model(s) may provide health economists with a tool for carrying out cost-utility analyses using clinical outcome data only.
This PDF is available to Subscribers Only