A Hybrid Cohort Individual Sampling Natural History Model of Age-Related Macular Degeneration: Assessing the Cost-Effectiveness of Screening Using Probabilistic Calibration

Abstract

Background. Age-related macular degeneration (AMD) is a leading cause of visual impairment and blindness. It is likely that treatment of AMD at earlier stages is more effective than later treatment; thus, screening for AMD should be considered. The aim of this study was to develop a natural history model of AMD to estimate the cost-effectiveness of screening. Methods. A hybrid cohort/individual sampling decision analytic model was developed. Primary data sets, expert elicitation, and data from the literature were used to populate the model. To incorporate joint parameter uncertainty, and to populate unobservable parameters, an innovative form of probabilistic calibration was applied to a range of output parameters. Results. In the reference case, annual screening from age 60 y is the most cost-effective option, although this is subject to high levels of uncertainty. Alternative, age-specific utility values show that screening is predicted to be less cost-effective, assuming interventions that reduce progression to wet AMD moderately improve the cost-effectiveness of screening, whereas the addition of anti—vascular endothelial growth factor therapy for juxtafoveal or subfoveal wet AMD lesions improves the cost-effectiveness of screening significantly. Conclusions. The extent of the uncertainty around the mean results, and the additional resources and possible reorganization of services required to implement screening, indicate that it may be preferable to reduce the level of uncertainty before implementing screening for AMD. Initial actions may be best targeted at assessing how routine data may be used to describe clinical presentation, a screening pilot study, and a secondary costing study.