Developing an algorithm to convert routine measures of vision into utility values for glaucoma.

Purpose: Measures of quality of life called utility values (UVs) are needed to deliver the most cost-effective health care for glaucoma patients. UVs are rarely measured in clinical research and practice whereas clinical outcomes such as visual field are routinely collected. The aim of this study was to develop an algorithm that calculates UVs directly from combinations of routine measures of binocular visual field, visual acuity, and contrast sensitivity.

Methods: A total of 132 outpatients with primary open angle glaucoma were recruited. The Time Trade-off (TTO) question was administered during face-to-face interviews. Binocular ETDRS logMAR visual acuity (VA(B)), binocular Pelli-Robson contrast sensitivity (CS(B)), and Humphrey 24-2 monocular visual field tests were performed on the same day. Integrated (binocular) visual field (IVF) scores were derived. Tobit regression analyses were used to model utility values based on combinations of IVF, VA(B), CS(B) and other controlling factors.

Results: UVs recorded for 123 cases correlated significantly with both clinical measures of binocular visual function (r = -0.47, IVF; r = -0.48, VA(B); r = 0.50, CS(B); P <0.0001) and measures of vision-specific quality of life (r = 0.54-0.6, P <0.0001). Two final models incorporate terms for IVF and VA(B), with or without living arrangements, and explain 22% and 31% of variation in utilities. CS(B) was not included in either model due to co-linearity between CS(B) and VA(B) confounding the models.

Conclusion: The models provide preliminary algorithms for predicting the expected UVs for glaucoma populations directly from clinical outcomes collected routinely in clinical practice.