One of the oldest patient communities on Twitter is characterized by the hashtag #bcsm, and it is a forum for breast cancer patients and survivors. This community has been hosting a weekly, moderated chat for the last three years. This paper describes work that analyzes the content of these chats and explores their effectiveness for the patients. The computational analysis compares the engagement, linguistic, and psychological facets of patients while tweeting during the chats and while tweeting out of the chats, and shows that there is a significant difference in behavior by the patients in the two modes. The result shows the effectiveness of social media chats for cancer patients for the purposes of information exchange and support.