Prevalence and patterns of anti-angiogenic treatment discontinuation in neovascular age-related macular degeneration—insights from health claims data

PURPOSE: To assess patterns and institutional variability of anti-angiogenic treatment discontinuation in neovascular age-related macular degeneration (nAMD) within Switzerland, and to examine the association of digital home monitoring with treatment persistence. METHODS: Retrospective observational analysis combined nationwide health insurance claims data (Helsana, n = 4840) and clinical data from a major tertiary centre (LUKS, n = 1932; nAMD subset n = 1387). Discontinuation was defined as no further injections for 31-365 days after the last intravitreal injection (IVI) and classified as supervised ( ≥ 1 follow-up visit in this period) or unsupervised (no follow-up visits in this period). Variability was evaluated among the 20 largest providers and across seven regions. Digital home monitoring's impact on non-persistence was assessed in the clinical cohort. RESULTS: Substantial heterogeneity was observed. Supervised discontinuation ranged from 20.7 to 45.7% (pooled proportion 32.0%, 95% confidence interval [CI] 30.2-33.9%) and unsupervised from 10.9 to 25.7% (pooled proportion 18.2%, 95%CI 16.8-19.8%) across centres. In the LUKS cohort, 1-year persistence was 93.7%, dropping to 72.7% by year 4, with unsupervised discontinuation rising from 3.0 to 12.5%. In 100 patients using voluntary digital home monitoring non-persistence was less likely (adjusted OR 0.59, 95%CI 0.35-0.98); among those who discontinued, home-monitoring users were more likely to remain under supervision (74% vs 54%). CONCLUSIONS: Treatment discontinuation in nAMD is frequent and demonstrates striking institutional and regional variability. The heterogeneity in discontinuation patterns suggests that treatment persistence may be modifiable through targeted, systematic, and patient-centred interventions. Digital home monitoring may be a promising strategy to support engagement and reduce non-persistence.