Understanding the preferences of younger women for the delivery of a service to predict breast cancer risk: a discrete choice experiment Short title: Preferences for a breast cancer risk-prediction service

Background: This study aimed to understand the preferences of a sample of younger women (30-39 years) for the attributes of models of service delivery for a breast cancer risk-prediction service to identify how best to design a service to optimise uptake.

Methods: A discrete choice experiment was used to quantify the preferences of a purposive sample of younger women (aged 30 to 39) without prior knowledge of their risk of developing breast cancer. Respondents chose from a series of questions including two unlabelled alternatives, representing different models of a risk-prediction service, and an opt-out alternative. Data were analysed using random parameter logit and latent class models to explore potential heterogeneity in preferences for the intervention.

Results: The predicted uptake for a risk-prediction service ranged from 77-89%. Participants preferred a service with more flexible appointments which could be booked by the individual themselves. Latent class analysis suggested that around 7% of women would never have their risk predicted and for approximately 30% of women the choice would depend on the design of the service.

Conclusion: Younger women would be likely to choose to have their breast cancer risk, although some groups were sensitive to the design of the prediction service.