User-Centered Design of Interactive Recommender Systems for Cancer Patients and Caregivers

User-centered design is a commonly accepted approach in designing interactive systems; however, its application to the design of interactive recommender systems is still limited. This paper reviews the design of H ELPeR, a patient-focused interactive recommender system for ovarian cancer ( O vCa) patients and caregivers. It has a dual goal: to present an example of an interactive recommender system in a health domain and to demonstrate a case of user-centered design of a recommender system in this challenging domain. The design of H ELPeR was informed by foundational studies of patients and caregivers and collaboration with health experts to understand the user base and their needs. Foundational studies revealed how users transition from seeking information to sharing advice, leading to a dual-mode system that provides recommendations for novice users and search functionality for experienced users. Health experts provided insights into challenges such as trajectory misalignment, multi-topic complexity, and contextual mismatches, guiding refinements to both interaction design and recommendation algorithms. The iterative design process resulted in a system that dynamically adapts to user trajectories and supports diverse user preferences through a multi-topic carousel and search interface. The evaluation demonstrated H ELPeR’s positive usability and performance outcomes, highlighting its ability to address the evolving needs of O vCa patients and caregivers.