Cost-effectiveness of autologous retinal pigment epithelium and choroid translocation in neovascular AMD

Abstract

Aim: To assess the cost-effectiveness of autologous retinal pigment epithelium and choroid translocation (PATCH) in neovascular age-related macular degeneration (AMD).

Methods: Visual acuity and complication rates of published patient series were used to determine the incremental utility of treatment for the patient. The utility data applied assume that the better eye was affected. Comparator was a meta-analysis of recent control groups, in which patients received best supportive care. To assess cost-effectiveness, costs per quality adjusted life years (QALYs) and costs of avoiding low vision ("legal blindness", i.e. ≤20/200) were calculated. Costs were based on a German sick fund perspective and in a scenario on US costs. Robustness of the model was investigated by univariate and probabilistic multivariate sensitivity analysis (PSA).

Results: Cost-utility analysis showed surgery to be the dominant ("cost-saving") strategy for Germany and for the US in both, cost-effectiveness and cost-utility analysis (costs per QALY). In the sensitivity analysis the intervention remained dominant or cost-effective in all scenarios investigated. Clinical outcomes and duration of modeling were the most influential factors in the sensitivity analyses.

Conclusion: Therapy of neovascular AMD by PATCH is a cost-effective treatment option for selected patients, who are not well suitable for other current treatment options.

Keywords: QALY; age-related macular degeneration; choroidal neovascularization; cost-utility analysis; macular surgery.

Figure 1. Model structure

A combined decision tree (for year 1) and Markov model (for year 2-n) was applied to model cost-effectiveness. Detailled clinical data and costing sources are described in the methods section. The (M) marks those nodes in the decision tree, where the Markov long-term model part starts. Patients may transition to death from any Markov state. The two Markov states [0.05-0.1] and <0.05 together represent “legal blindness” as used in cost-effectiveness modeling

Figure 2. Probabilistic sensitivity analysis

The graph gives the likelihood of the surgical intervention to be below a given threshold for cost-effectiveness. It can be seen, that therapy is cost-effective even for low thresholds, e.g. >=90% probability for 6'874 �/QALY