The development of online social platforms has promoted the improvement of online health communities (OHCs). However, OHCs often ignore differences in user discussions caused by the characteristics of diseases. The purpose of this research was to study differences in the topics and emotions of patients with physiological and psychological diseases by mining the text that they posted in OHCs as well as to discuss how to satisfy these differences. The data came from Baidu Post Bar, the world's biggest Chinese forum. We collected 50,230 posts from heart disease, hypertension, depression and obsessive-compulsive bars. Then, we used topic modeling and sentiment analysis techniques on these posts. The results indicate that there are significant differences in the preferences of discussion and emotion between patients with physiological and psychological diseases. First, people with physiological diseases are more likely to discuss treatment of their illness, while people with psychological diseases are more likely to discuss feelings and living conditions. Second, psychological disease patients' posts included more extreme and negative emotions than those of physiological disease patients. These results are helpful for society to provide accurate medical assistance based on disease type to different patients, perfecting the national medical service system.
Keywords: disease difference; online health community; sentiment analysis; topic modeling.