People with diabetes who are insulin treated are at risk of experiencing hypoglycemia – a condition of potentially dangerously
low blood glucose levels. Predictive algorithms can aid these people in preventing hypoglycemia. In this two-part study we
applied user-centered design to explore how a personalized nocturnal hypoglycemia risk prediction can be communicated in
a mobile self-management app (SMA) for people with type 2 diabetes. In a preliminary understanding study, we conducted
semi-structured interviews with potential users (N=10) and one endocrinologist to acquire domain knowledge, establish user
needs, and identify preferences for visual presentation of hypoglycemia risk. Based on our initial findings, we developed a
mobile application prototype which was evaluated in a think-aloud evaluation with prospective users (N=5) to validate
user needs and verify design and functionality. Our findings generally align with previous research in emphasizing the
significance of using color cues and simple visualizations supplemented with concise text or numbers. We also found that
the SMA should provide flexibility as users need different levels of detail and that the user should be able to personalize the
prediction configuration. We further identified barriers and enablers for adoption of mobile diabetes SMA and highlight
the importance of collaboration between users, health professionals, and researchers. Based on our findings we draw
recommendations for future work in this field.