Santarossa Sara, Rapp Ashley, Sardinas Saily, Hussein Janine, Ramirez Alex, Cassidy-Bushrow Andrea E, Cheng Philip, Yu Eunice
Department of Public Health Sciences Henry Ford Health System Detroit, MI United States.
School of Medicine Wayne State University Detroit, MI United States.
JMIR Infodemiology. 2022 Feb 22;2(1):e31259. doi: 10.2196/31259. eCollection 2022 Jan-Jun.
The scientific community is just beginning to uncover the potential long-term effects of COVID-19, and one way to start gathering information is by examining the present discourse on the topic. The conversation about long COVID-19 on Twitter provides insight into related public perception and personal experiences.
The aim of this study was to investigate the #longCOVID and #longhaulers conversations on Twitter by examining the combined effects of topic discussion and social network analysis for discovery on long COVID-19.
A multipronged approach was used to analyze data (N=2500 records from Twitter) about long COVID-19 and from people experiencing long COVID-19. A text analysis was performed by both human coders and Netlytic, a cloud-based text and social networks analyzer. The social network analysis generated Name and Chain networks that showed connections and interactions between Twitter users.
Among the 2010 tweets about long COVID-19 and 490 tweets by COVID-19 long haulers, 30,923 and 7817 unique words were found, respectively. For both conversation types, "#longcovid" and "covid" were the most frequently mentioned words; however, through visually inspecting the data, words relevant to having long COVID-19 (ie, symptoms, fatigue, pain) were more prominent in tweets by COVID-19 long haulers. When discussing long COVID-19, the most prominent frames were "support" (1090/1931, 56.45%) and "research" (435/1931, 22.53%). In COVID-19 long haulers conversations, "symptoms" (297/483, 61.5%) and "building a community" (152/483, 31.5%) were the most prominent frames. The social network analysis revealed that for both tweets about long COVID-19 and tweets by COVID-19 long haulers, networks are highly decentralized, fragmented, and loosely connected.
This study provides a glimpse into the ways long COVID-19 is framed by social network users. Understanding these perspectives may help generate future patient-centered research questions.
科学界刚刚开始揭示新冠病毒病（COVID-19）的潜在长期影响，而开始收集信息的一种方法是审视当前关于该主题的论述。推特上关于新冠后长期症状（long COVID-19）的讨论提供了对相关公众认知和个人经历的洞察。
本研究的目的是通过检查主题讨论和社交网络分析对发现新冠后长期症状的综合影响，来调查推特上关于#longCOVID和#longhaulers的讨论。
采用多管齐下的方法来分析关于新冠后长期症状以及来自经历新冠后长期症状的人的数据（来自推特的N = 2500条记录）。由人工编码员和Netlytic（一个基于云的文本和社交网络分析器）进行文本分析。社交网络分析生成了显示推特用户之间联系和互动的姓名网络和链条网络。
在2010条关于新冠后长期症状的推文和490条由新冠后长期症状患者发布的推文中，分别发现了30923个和7817个独特词汇。对于这两种对话类型，“#longcovid”和“covid”是提及最频繁的词汇；然而，通过直观检查数据，与患有新冠后长期症状相关的词汇（即症状、疲劳、疼痛）在新冠后长期症状患者发布的推文中更为突出。在讨论新冠后长期症状时，最突出的框架是“支持”（1090/1931，56.45%）和“研究”（435/1931，22.53%）。在新冠后长期症状患者的对话中，“症状”（297/483，61.5%）和“建立社区”（152/483，31.5%）是最突出的框架。社交网络分析显示，对于关于新冠后长期症状的推文和新冠后长期症状患者发布的推文，网络都是高度分散、碎片化且联系松散的。
本研究让我们得以一窥社交网络用户对新冠后长期症状的构建方式。理解这些观点可能有助于提出未来以患者为中心的研究问题。
