Background: The spreading of health-related rumors can profoundly put society at risk, and the investigation of strategies and methods can efficiently prevent the dissemination of hazardous rumor is necessary, especially during a public health emergency including disease outbreaks. In this article we review the studies that implicated the surveillance system in identifying rumors and discuss the different aspects of current methods in this field. Methods: We searched PubMed, EMBASE, Scopus, and Web of Science databases for relevant publications in English from 2000 to 2020. The PICOS approach was used to select articles, and two reviewers extracted the data. Findings were categorized as a source of rumors, type of systems, data collection, and data transmission methods. The quality of the articles was assessed using the Mixed Method Appraisal Tool (MMAT) checklist. Results: Five studies that presented the methods used for rumor detection in different outbreaks were included in the critical appraisal process. Findings were grouped into four categories: source of rumors, type of systems, data collection, and data transmission methods. The source of rumors in most studies was media, including new social and traditional media. The most used data collection methods were human-computer interaction technique, and automatic and manual methods each were discussed in one study. Also, the data transmission method was asynchronous in the majority of studies. Conclusion: Based on our findings, the most common rumor detection systems used in the outbreaks were manual and/or human-computer methods which are considered to be timeconsuming processes. Due to the ever-increasing amount of modern social media platforms and the fast-spreading of misinformation in the times of outbreaks, developing the automatically and real-time tools for rumor detection is a vital need.