Objective   To examine the impact of risk information characteristics on the Chinese public's risk perception of coronavirus disease 2019 (COVID-19) and to put forward targeted suggestions for improving infectious disease epidemic information release.

  Methods  Using a self-designed questionnaire, we conducted an online anonymous survey among 1 446 Chinese internet users recruited with snowball sampling in five provincial-level regions of China from June 15 to 18, 2020.

  Results   Valid information were collected from 1 292 participants (43.0% males and 56.0% females) averagely aged 30.4 ± 9.3 years. Evaluated by a scale with the highest score of 5 for the greatest impact, the scores for the impact of characteristics of COVID-19 related information on participants' the epidemic risk perception were 3.37 ± 0.98 for personal experience associated with infectious disease, 3.81 ± 1.21/3.61 ± 0.79/3.48 ± 1.13/3.52 ± 0.90/2.92 ± 0.95 for obtaining information disseminated by government departments/experts or scholars/traditional media/internet or social media/neighbors or friends, 4.36 ± 0.59 for inconsistent or contradictory information, 4.35 ± 0.61 for different or controversial information from experts or scholars, 3.74 ± 0.87 for hearing about discussions or disputes among opinion leaders such as very important Microblog users or stars, 3.99 ± 0.94 for exposed to a great amount of information about COVID-19 during the epidemic, 3.75 ± 1.00 for rarely accessing to information other than COVID-19 epidemic, 3.53 ± 1.05 for obtaining information about previous infectious disease epidemic such as severe acute respiratory syndrome and influenza, 3.57 ± 0.94 for accessing information about outbreaks of other infectious diseases in foreign countries, 2.50 ± 1.08 for obtaining information on phenomenon or problems induced by the epidemic, 3.36 ± 0.98 for accessing text descriptions/photographs/videos on severe harms caused by the epidemic, and 3.48 ± 1.13 for accessing the information on panic or sadness induced by the epidemic among the public. The 5 common factors identified through factor analysis were source, dispute, quantity, association, and sentiment tendency of information accessed, with the corresponding variance contribution rate of 17.334%, 13.638%, 13.221%, 13.111%, and 10.593%, respectively, and a cumulative variance contribution rate of 67.897% for the 5 common factors.

  Conclusion  The source, dispute, quantity, association, and sentiment tendency are main characteristics of COVID-19-related information affecting the public′s disease risk perception and information source is a primary factor among the five information characteristics.