Abstract: 【Purpose/significance】The sustainable development of online healthy communities lies in knowledge co-creation. As the
starting point of knowledge co-creation, users' information demands are an important basis for the research on the mechanism of
knowledge co-creation in online health communities.【Method/process】Firstly, the medical treatment situations of users in online
health communities are described from the three dimensions of medical treatment channels, medical treatment status, and treatment
status, and four types of medical treatment situations are proposed. Secondly, 97,169 valid question data of users in the 2020 full-year
"Question and Answer" community of the Medical Questionnaire.com are collected by using python, and each question data is labeled
with medical treatment situation labels by independent coding. Finally, Word2vec and LDA algorithms are combined to extract users'
information demand preferences in different medical treatment situations.【Result/conclusion】It is found that 30.1% of users could
not accurately describe their medical treatment situation. Among the users who accurately expressed the medical situation, the online
initial diagnosis situation is the main medical situation, mainly focusing on symptoms and medical treatment; offline initial diagnosis +
no treatment + online further consultation mainly focuses on examination and causes; offline initial diagnosis + treatment + online fur?
ther consultation mainly focuses on the method effect and follow-up; online initial diagnosis + treatment + online further consultation
mainly focuses on method effects and follow-up.【Innovation/limitation】From the whole process of disease consultation and treatment,
this paper proposes four types of medical treatment situations, and reveals user' information demand preferences under different medi?
cal treatment situations. However, only the doctor-patient type of online health community is empirically analyzed.