探討新冠肺炎(COVID-19)之確診率、死亡率及復原率之可能相關因素

在新冠肺炎疫情大流行之下，全球已累計破億例確診個案，其中更有數百萬人死亡，已演變成一場全球性大災難。本研究依時間的橫斷面與時間的縱深方向兩方面進行資料分析。前者，採用獨立樣本的廣義線性模型(GLM)來探討新冠肺炎(COVID-19)的確診率(感染率)、死亡率及復原率的可能影響因素；後者，我們利用網路爬蟲程式將六大洲(Continents)共184個國家，從2020/01/22至2021/02/28每天感染人數及死亡人數摘錄下來。每一國家重複登錄404筆資料，屬於重複測量的Dependent data。因此，我們分別對六大洲使用GEE方法的GLM來比較，經調整時間與人口總數效應後，各大洲前十大感染心冠肺炎風險及感染速率的國家。從獨立樣本的GLM來看，在確診率的部分卡介苗、國家之國民所得、及六大洲皆為顯著影響因素；在死亡率的部分，卡介苗及國家之國民所得為顯著因素；在復原率的部分，卡介苗、國家之國民所得、及六大洲皆為顯著因素。從GEE方法的廣義線性模式之結果可知: (1)以歐洲來說，感染風險最高為安道爾共和國，死亡風險最高的為義大利；(2)以北美洲來說，感染風險最高的為美國，死亡風險最高的為墨西哥；(3)以亞洲來說，感染風險最高的為卡達，死亡風險最高的為葉門；(4)以南美洲來說，感染風險最高的為法屬圭亞那，死亡風險最高的為厄瓜多；(5)以非洲來說，感染風險最高的為馬約特，死亡風險最高的為厄利垂亞；(6)以大洋洲來說，感染風險最高的為法屬玻里尼西亞，死亡風險最高的為澳洲。

關鍵字

新冠肺炎；廣義線性模型

並列摘要

Under the pandemic of COVID-19, there have been more than 100 million confirmed cases worldwide, among them millions of people have died. This has turned into a global disaster. In this study, data was collected in both the time cross-sectional method and the longitudinal method. In the former, the independent sample generalized linear models (GLM) were used to explore the possible prognostic factors of the infection rate, mortality rate and recovery rate of the COVID-19; in the latter, we have a total of 184 countries from six continents. The daily cumulative total number of infections and deaths from 2020/01/22 to 2021/02/28 in each country was extracted using a web crawler program. For each country, there are 404 repeated measurements in total. Accordingly, for each continent, we use the GEE methods’ GLM to compare the top ten countries with the highest risk of COVID-19 infections and the highest infection rates, after adjusting for the effects of time and population size. The results of independent sample GLM showed that: (1) For the infection rate, BCG vaccination, income level of the country and continents are significant factors; (2) For the mortality rate, BCG vaccination and income level of the country are significant factors; (3) For the recovery rate, BCG, income level of the country and continents are significant factors. The results of the GEE method’s GLM showed that: (1) For Europe, the highest infection rate is Andorra, and the highest mortality rate is Italy; (2) For North America, the highest infection rate is USA, and the highest mortality rate is Mexico; (3) For Asia, the highest risk of infection is Qatar, and the highest mortality rate is Yemen; (4) For South America, the highest infection rate is French Guiana, and the highest mortality rate is Ecuador; (5) For Africa, the highest risk of infection is Mayotte, and the highest mortality rate is Eritrea; (6) For Oceania, the highest infection rate is the French Polynesia, and Australia has the highest mortality rate.