Clustering techniques for organizing cancer-related concepts into meaningful groups for patients.

A major challenge in the design of Internet support for patients is to organize the content in a way that patients perceive as meaningful and easy to find. We used card sorting and clustering techniques to understand how patients organize cancer-related symptoms and problems into meaningful groups, and compared their symptom categorizations to those done by professional nurses. Patients sorted symptoms into significantly fewer groups than nurses, and had difficulties in distinguishing between symptom types and their associations. Nurses typically organized symptoms by physiological or physical body structure, and classified problems around physical, emotional, and social functioning. This study shows how patients categorize symptoms and problems differently from the professionally-oriented conceptual structures of disease. We used this information to improve the design of a cancer patient support system.