Measuring and valuing glaucoma health states

Abstract

Cost-utility is a method that integrates clinical factors of disease progression with cost and consequence to patient’s overall health state. Cost-utility analyses describe the trade-offs that are consequences of interventions and are commonly used to support health policy decision making. Well constructed cost-utility analyses can improve access patients have to effective interventions and move past ineffective or exceedingly costly interventions. The methods employed to assess incremental effectiveness of glaucoma interventions have advanced over time, where investigators have adapted to standards for reporting and modeling strategies. However, most cost-utility analyses use a restrictive set of utility values from foundational, but outdated work. Glaucoma cost-utility analyses combine utility values from patient preferences, on multiple scales of measurement. These methodological flaws in the use and reporting of utility values limit the internal validity of the study and external validity in generalizing the results to broader populations. Common generic preference-based measures of health utility, including the EuroQol 5 Dimension 3/5 Level (EQ-5D-3/5L), lack the sensitivity to detect health state changes in patients with glaucoma. The three main investigations of this dissertation aimed to describe the current methodological issues in glaucoma cost-utility investigations, validate a condition-specific preferencebased measure (HUG-5) in a sample of patients with various stages of glaucoma, and generate a preference-based weighting algorithm for glaucoma health states for US general population. A systematic review of cost-utility analyses in the literature revealed that most all cost-utility analyses failed to document rationale for including health utility estimates or conduct a review of appropriate utilities for relevant health states. Most authors of cost-utility analyses did not recognize or describe the limitations of combining utilities from multiple scales of measurement and from different preference sources. In a psychometric validation of the HUG-5, a sample of patients with glaucoma completed the NEI-VFQ-25, the HUG-5, and the EQ-5D-5L to describe their quality of life and health state. The HUG-5 dimensions described similar dimensions measured by National Eye Institute Visual Function Questionnaire (NEI-VFQ-25) scales and differentiated from unrelated dimensions with adequate test-retest reliability on 3 month follow up. To model utility values to HUG-5 health states, a multi-attribute disutility function (MADUF) was estimated from the preferences of the US general population. A preference-weighting algorithm assigned utility values to the 3125 health states measured by the HUG-5. In validating the MADUF with directly elicited marker health states, the MADUF performed similar in mean absolute error relative to other related studies for the Health Utilities Index Mark 3 (HUI-3) and Patient-Oriented Prostate Utility Scale (PORPUS). This work has significant implications for investigators evaluating glaucoma interventions with cost-utility comparisons, providing an alternative measure of utility on the policy scale with a condition-specific preference-based measure.