Abstract
Two complement inhibitors are currently used for treating geographic atrophy (GA) secondary to age-related macular degeneration (AMD) and other treatments are in development. We surveyed people with GA to better understand how efficacy, risks, and intravitreal injection frequency impact their treatment preferences. We used a mixed-methods approach (including qualitative research, pretesting interviews, and clinical expert and patient input) to design an online preference elicitation survey accessible to people with visual impairment. Preferences of US-based adults with GA were elicited with multi-dimensional thresholding. Participants ranked the included treatment characteristics and completed four sets of trade-off exercises (with two to three questions per exercise) containing two characteristics per question. The relative importance of treatment characteristics and participants' willingness to trade-off between characteristics were evaluated. Participants (N=77; mean age 69.3 years; 32.5% GA treatment-experienced) often preferred treatment over no treatment, ranging from 57% preferring treatment for the "worst" assessed treatment to 99% for the "best" assessed. Participants' decisions were most influenced by extending the time until they were unable to read (relative importance: 32%) or time to see half as far (27%), followed by injection frequency (16%) and reducing risks of developing wet AMD (14%) and eye inflammation (12%). Most participants (86%) preferred a one-time intravitreal injection (vs every other month or every month). Within an 18-month time treatment timeframe, on average, participants were willing to accept higher risks for increased efficacy (eg a 35% higher risk of eye inflammation or 29% higher risk of wet AMD for a treatment that increases efficacy from 1 to 4 years longer to read and to see half as far). All else equal, participants were willing to accept higher risks for a one-time injection over monthly injections (eg 14% higher risk of eye inflammation or 12% higher risk of wet AMD). Most participants preferred the "worst" treatment profile over no treatment, highlighting a potential unmet need for GA treatment. Treatment choices were most influenced by efficacy, particularly prolonging the time until they are unable to read. Participants were willing to tolerate significantly higher risks for their preferred injection frequency. Shared treatment decision-making is encouraged to ensure alignment with patients' individual preferences.