Abstract: 【Purpose/significance】With the huge scale of online health community users and a vast amount of information,how to help
community managers and users to discern useful information and improve the efficiency of decision making is an urgent problem.
【Method/process】A new comment usefulness analysis framework is proposed under the perspective of complex networks.First,collect the data of patients' comments in online health communities，use text analysis to analyze the thematic and emotional distribution of useful comments,non-useful comments and all comments，and analyze the usefulness characteristics of various types of comment texts preliminarily; secondly,transform all kinds of comment texts into text association networks,and use social network analysis method to analyze their usefulness ulteriorly; finally,analyze the correlation between the usefulness of comments,their characteristics and the pa? tient comments,user votes on the usefulness of the comments,structural features of text association network to implement review useful? ness analysis based on text association network【. Result/conclusion】There are some differences in the text association network struc? ture between useful and non-useful comments,and the focus of information needs and experience output of users before and after con? sultation in online health communities is different.【Innovation/limitation】Based on a complex network perspective,study the useful? ness of online health community comments，but only use the data from haodf.com.To study the usefulness of information content of a larger number and variety of online health communities in the future.