[Purpose/significance] The aim is to understand the content characteristics and quality of health rumor-refuting information in the platform, and optimize the platform’s service and rumor refuting effectiveness accordingly.
[Method/process] Taking the "Scientific rumor refuting platform" as an example, we follow the logical chain of "whathow
is it-how to do" to carry out a systematic analysis of the health rumor-refuting information on the platform. First of
all, the LDA topic model is applied to mine the content topics and evolutionary trends of health rumor-refuting information. Secondly, combined with the quality evaluation framework of DISCERN and Michigan Checklist, we evaluate the quality of health rumor- refuting information from multiple dimensions such as attribution, timeliness. And we use ANOVA to explore the relationship between content elements and the quality evaluation results. Finally, we propose the management measures to optimize the platform construction. [Result/conclusion] Five topics such as nutritional value and food additives are the main topics of health rumor-refuting information. The information contents of the platform present obvious thematic evolution and intensity changes. The overall quality of health rumor-refuting information on the platform is not high, and the information quality is closely related to the content theme, expert information, reference literature and other content elements. In the future, the platform can optimize the health rumor-refuting services based on the above conclusions, so as to help create a good online health information ecology.