Abstract: 【Purpose/significance】Online health community has become an important form for the public to obtain medical information
and services.This paper explores the identification of key users and their characteristics in online health community,with a view to pro‐
vide theoretical basis for improving the quality and efficiency of health community service【. Method/process】Based on the information behavior theory.a multi-dimensional analysis framework was constructed.including interactive behavior attributes,information quality attributes and emotional tendency attributes.AttriRank algorithm and network invulnerability evaluation method were used to identify key users of online health community【. Result/conclusion】The results showed that 15 key users were identified in the QQ group of bili‐ary cancer diseases and they not only have the characteristics of high activity and reciprocity.but also have the characteristics of highdiversity and balanced information quality.and most of them have positive emotional tendencies.The analysis framework of "behavior +content + emotion" and the user ranking algorithm considering attributes can accurately identify the key users of the online health com‐munity.which provides scientific decision support for the sustainable operation of online health community.【Innovation/limitation】Build a multi-dimensional attribute analysis framework for online health community key user identification,which enriches the theoret‐ical system of online health community key user identification。