Department of Pathophysiology, Faculty of Basic Medical Sciences, Guilin Medical University, Guilin, Guangxi, China.
Department of Environmental Health and Occupational Medicine, School of Public Health, Guilin Medical University, Guilin, Guangxi, China.
J Med Virol. 2020 Sep;92(9):1587-1595. doi: 10.1002/jmv.25834. Epub 2020 Apr 25.
This study seeks to examine and analyze the spatial and temporal patterns of 2019 novel coronavirus disease (COVID-19) outbreaks and identify the spatiotemporal distribution characteristics and changing trends of cases. Hence, local outlier analysis and emerging spatiotemporal hot spot analysis were performed to analyze the spatiotemporal clustering pattern and cold/hot spot trends of COVID-19 cases based on space-time cube during the period from 23 January 2020 to 24 February 2020. The main findings are as follows: (1) The outbreak had spread rapidly throughout the country within a short time and the current totality incidence rate has decreased. (2) The spatiotemporal distribution of cases was uneven. In terms of the spatiotemporal clustering pattern, Wuhan and Shiyan city were the center as both cities had high-high clustering pattern with a surrounding unstable multiple-type pattern in partial areas of Henan, Anhui, Jiangxi, and Hunan provinces, and Chongqing city. Those regions are continuously in the hot spot on the spatiotemporal tendency. (3) The spatiotemporal analysis technology based on the space-time cube can analyze comprehensively the spatiotemporal pattern of epidemiological data and produce a visual output of the consequences, which can reflect intuitively the distribution and trend of data in space-time. Therefore, the Chinese government should strengthen the prevention and control efforts in a targeted manner to cope with a highly changeable situation.
本研究旨在探讨和分析 2019 年新型冠状病毒病（COVID-19）疫情的时空模式，并确定病例的时空分布特征和变化趋势。因此，基于时空立方体，我们采用局部异常值分析和新兴时空热点分析方法，对 2020 年 1 月 23 日至 2 月 24 日期间 COVID-19 病例的时空聚类模式和冷/热点趋势进行了分析。主要发现如下：（1）疫情在短时间内迅速蔓延至全国，目前总发病率有所下降。（2）病例的时空分布不均匀。在时空聚类模式方面，武汉和十堰市是中心城市，均呈高高聚类模式，而河南省、安徽省、江西省和湖南省以及重庆市部分地区呈周围不稳定多型模式，这些地区在时空趋势上持续处于热点状态。（3）基于时空立方体的时空分析技术可以全面分析疫情数据的时空模式，并产生结果的可视化输出，直观地反映数据在时空上的分布和趋势。因此，中国政府应加强有针对性的防控措施，应对高度变化的形势。
