Demographic change in regard to an aging population with an increasing amount of diabetes patients will put a strain on health care rentability in all modern societies. Electronic living assistants for diabetes patients might help lift the burden on taxpayers, if they are usable for the heterogeneous user group. Research has shown that correct mental models of device menu structures might help users in handling electronic devices. This exploratory study investigates construction and facilitation of spatial mental models for a menu structure of a diabetes living assistant and relates them to performance in usage of a device. Furthemore impact of age, domain knowledge and technical expertise on complexity and quality of the mental model are evaluated. Results indicate that even having a simplified spatial representation of the menu structure increases navigation performance. Interestingly not the overall correctness of the model was important for task success but rather the amount of route knowledge within the model.