The Characteristics of Doctors, Patients and Question Records in Online Medical Communities

Objective: This article starts to explore the characteristics of users, doctors and the question & answer records related to childhood asthma, by analyzing the data from three online medical communities. It then develops a topic model to identify the topics in the question-answer records which can help to find specific needs of users.

Method: We developed a topic model to identify the topics in the question-answer records which can help to find specific needs of users. After data cleaning and word segmentation, we applied a text mining technique, Latent Dirichlet Allocation (LDA) topic modeling, to analyze the themes of the question records and the characteristics of patients and doctors.

Results: Finally, 2173 question records were divided into eight themes which included knowledge requirements, emotional needs and nursing needs. Network data were found to correspond with childhood asthma epidemiology. Fewer doctors from children’s speciality hospitals were engaged with the online medical communities than other types of facilities.

Conclusion: The data in online medical communities can reflect the needs of users, which can help medical staff to improve medical services.

Keywords: online medical communities, Latent Dirichlet Allocation, thematic analysis, childhood, asthma, health information needs.