Purpose:
Most previous utilities work in the visual sciences has used perfect vision as the upper anchor. However, common cost-effectiveness standards were derived using perfect health-anchored utilities. Applying these standards to perfect vision-anchored utilities requires the assumption that the visual component of perfect health is additive with non-visual components, but there is no evidence to support this.
Methods:
Patients with exactly one of the conditions of interest were recruited from academic practices.All participants completed the NEI-VFQ and SF-36. Using the computer program U-Titer-II, standard gamble utilities were measured to assess multiple scenarios: (1) given current health and vision, willingness to risk instant death for perfect health including perfect vision, (2) given current vision, willingness to risk instant unilateral blindness for perfect vision, and (3) given current health but perfect vision, willingness to risk instant death for perfect health. Spearman correlations between VFQ and SF-36 scales were also calculated.
Results:
Scenario #2 (perfect vision) utilities were lower than Scenario #1 (perfect health) ones for the same participants, particularly with increased disease severity. Scenario #3 (comorbidity) utilities were similar to Scenario #1, indicating that little of the overall perceived loss of quality of life comes from visual disease if the scales are additive. Correlations between VFQ and SF-36 were weak for all domains other than those related to general health.
Conclusions:
If visual health were an additive component of perfect health, then the similarity of Scenarios #1 and 3 would mean that vision is an insignificant part of overall health. Scenario #2, patient behavior, and the modesty of the correlation between VFQ and SF-36 domains suggests instead that the conceptual relationship between vision and general health is complex and requires additional investigation.
Keywords: quality of life • clinical research methodology