Department of Gastrointestinal Hepatobiliary, Chi Mei Jiali Hospital, Tainan 700, Taiwan.
Department of Medical Research, Chi-Mei Hospital, Tainan 700, Taiwan.
Int J Environ Res Public Health. 2021 Mar 3;18(5):2461. doi: 10.3390/ijerph18052461.
The COVID-19 pandemic has spread widely around the world. Many mathematical models have been proposed to investigate the inflection point (IP) and the spread pattern of COVID-19. However, no researchers have applied social network analysis (SNA) to cluster their characteristics. We aimed to illustrate the use of SNA to identify the spread clusters of COVID-19. Cumulative numbers of infected cases (CNICs) in countries/regions were downloaded from GitHub. The CNIC patterns were extracted from SNA based on CNICs between countries/regions. The item response model (IRT) was applied to create a general predictive model for each country/region. The IP days were obtained from the IRT model. The location parameters in continents, China, and the United States were compared. The results showed that (1) three clusters (255, n = 51, 130, and 74 in patterns from Eastern Asia and Europe to America) were separated using SNA, (2) China had a shorter mean IP and smaller mean location parameter than other counterparts, and (3) an online dashboard was used to display the clusters along with IP days for each country/region. Spatiotemporal spread patterns can be clustered using SNA and correlation coefficients (CCs). A dashboard with spread clusters and IP days is recommended to epidemiologists and researchers and is not limited to the COVID-19 pandemic.
新冠疫情在全球范围内广泛传播。许多数学模型已经被提出，以研究新冠疫情的拐点（IP）和传播模式。然而，没有研究人员应用社会网络分析（SNA）对其特征进行聚类。我们旨在说明如何应用 SNA 来识别新冠疫情的传播集群。从 GitHub 下载了国家/地区的累计感染病例数（CNIC）。基于国家/地区之间的 CNIC，从 SNA 中提取了 CNIC 模式。应用项目反应模型（IRT）为每个国家/地区创建了一个通用预测模型。从 IRT 模型中获得了 IP 日。比较了各大洲、中国和美国的位置参数。结果表明：(1) 使用 SNA 可以分离出三个集群（255、n = 51、130 和 74，模式从东亚和欧洲到美洲）；(2) 中国的平均 IP 日数较短，位置参数较小；(3) 利用在线仪表板可以显示每个国家/地区的集群及其 IP 日数。使用 SNA 和相关系数（CC）可以对时空传播模式进行聚类。建议为流行病学家和研究人员提供带有传播集群和 IP 日数的仪表板，且不限于新冠疫情。
