Cluster analysis on epidemic status of COVID-19 among different provinces

Abstract:

Objective：To give the comprehensive evaluation on epidemic status of coronavirus disease 2019（COVID-19） among dif－ferent provinces and classify all provinces into reasonable classes in accordance with the severity. Methods：Systematic clustering method was used to classify the development trend of cumulative confirmed cases，cumulative death cases and cumulative cured cases. Clustering heat maps were used for visual display and analysis. Results：The first class included Hubei. A total of 56 249 con－firmed cases in Hubei were much more than the total of 12 335 cases in other regions. The second class included Henan. Confirmed cases of 1 321 in Henan ranked to the third in the country and the death number was the largest（13 cases） except Hubei. The third class included Jiangxi，Zhejiang，Jiangsu，Shanghai，Sichuan，Anhui，Hunan，and Guangdong. These provinces had a higher confirmed cases but lower mortality compared with the following classes. The fourth class included Shandong，Chongqing，Beijing，Guangxi，Gansu，Tianjin，Heilongjiang，Hainan and Hebei. These provinces had a median confirmed patients，cure rate and mortality. The fifth class included Shanxi，Yunnan，Fujian，Shaanxi，Liaoning，Guizhou and Jilin. These provinces had the a lower confirmed patients and very few death cases（only 3 cases in this class）. The sixth class included Tibet，Taiwan，Qinghai，Ningxia，Macau，Inner Mongolia，Hong Kong and Xinjiang. These provinces had few confirmed cases（less than 71 cases） and only 2 death cases. Conclusion：In summary，areas that need to be closely controlled are Hubei and Henan，where can be regarded as extremely high-endemic areas. Areas that need to be closely monitored are Jiangxi，Zhejiang，Jiangsu，Shanghai，Sichuan，Anhui，Hunan and Guangdong，where can be regarded as high-endemic areas. Tibet，Taiwan，Qinghai，Ningxia，Macau，Inner Mongolia，Hong Kong and Xinjiang are relatively safe and can be regarded as low-endemic areas，while other areas are moderately endemic areas.