新型冠狀病毒(COVID-19)流行初期確診率與死亡率的相關因子：以全球空間資料分析
No Thumbnail Available
Date
2022
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
背景：
新型冠狀病毒 (Coronavirus disease 2019, COVID-19)在2019年12月於中國湖北省發現多起群聚感染，且迅速擴散至全中國並蔓延至其他國家，造成大流行疾病，臺灣在2020年1月21日，出現第1例COVID-19境外移入，COVID-19除了對經濟造成衝擊外，也因各國疫情及病例數與日俱增，讓民眾對於未知的疾病產生恐慌，本研究目的在了解世界各國COVID-19流行初期確診率及死亡率分佈情形及相關因子。
研究方法：
本研究透過世界各國的開放式數據探討眾多與COVID-19確診率及死亡率相關的因子，如：肥胖、高齡化、平均壽命、經濟發展程度、識字率、人口密度、傳染病與慢性病(癌症、心血管疾病、糖尿病、肺結核等的盛行率)、基礎衛生建設涵蓋率、醫療資源(醫師密度、病床密度)。空間統計分析用於探索 COVID-19確診率和死亡率的空間分佈。 本研究按兩個指標日期進行空間統計分析，分別是第一個分析時間點（2020年7月15日，T1）和第二個分析時間點（2020年12月15日，T2）。為了探索相關的因子和結果變項之間的空間關聯，我們進行了不同定義的空間相關矩陣之空間自迴歸分析，包括一階國界相鄰、二階國界相鄰、500公里距離相鄰、1,000 公里距離相鄰和 1,500公里距離相鄰。空間自迴歸分析並考量自變項共線性的問題。研究結果：
共175個國家的資料納入空間統計分析，以1,500公里距離相鄰為定義，排除部份共線性的自變項後，將剩餘的自變項同時納入多變數空間自迴歸分析顯示，國內生產毛額 (單位：每壹美元，估計係數=0.46，p=0.033)及肥胖率 (單位：每100人，估計係數=0.95, p<0.001)與T1確診率 (每百萬人)有顯著正相關；國內生產毛額 (單位：每壹美元，估計係數=0.43，p=0.029)及肥胖率 (單位：每100人，估計係數=1.02, p<0.001)也與T2確診率(每百萬人)有顯著正相關。
以1,500公里距離相鄰為定義，排除部份共線性的自變項後，將剩餘的自變項同時納入多變數空間自迴歸分析顯示，肥胖率 (單位：每100人，估計係數=0.75,p=0.001)及女性肺癌死亡率 (單位：每十萬人，估計係數=0.65, p=0.047)與T1死亡率 (每百萬人)有顯著正相關；僅肥胖率 (單位：每100人，估計係數=0.93, p<0.001)與T2死亡率 (每百萬人)有顯著正相關。
結論：
本研究結果與過去研究相符。在控制潛在的干擾變項後，研究結果顯示肥胖率與COVID-19確診率及死亡率有明顯的正相關，其次經濟與慢性疾病也是確診率的重要相關因子。關鍵字：新型冠狀病毒、大流行疾病、多變數空間自迴歸分析、確診率、死亡率
Background： Coronavirus disease 2019 (COVID-19) was discovered in China in December 2019, and it quickly spread to the whole of China and other countries, leading to a pandemic. The first case of COVID-19 from overseas immigration occurred in Taiwan on January 21, 2020. Besides its impact on the economy, COVID-19 has also caused public panic about unknown diseases due to the pandemic in many countries and the increasing number of cases worldwide. The purpose of this study was to investigate theCOVID-19 case rate and mortality in the early stage of the pandemic and reveal associated factors by analyzing countries' data worldwide. Method： Our study used open data of countries around the world to explore the associated factors with the COVID-19 case rate and mortality, including obesity, aging, life expectancy, economic development index, literacy rate, population density, infectious diseases, and chronic diseases (the prevalence of cancer, cardiovascular disease, diabetes, tuberculosis), basic sanitation rate, medical resources (physician density, hospital bed density) were adopted by research. Spatial autoregression models were used to explore the spatial distribution of the COVID-19 case rate and mortality. The study performed spatial autoregression analysis by two index dates, the first analysis time point (July 15, 2020, T1) and the second analysis time point (December 15, 2020, T2), respectively. To explore the spatial association between explanatory variables and outcomes, spatial autoregression models by a different definition of the spatial-correlation matrix were performed, including first-order adjacent border, second-order adjacent border, 500 kilometers adjacent, 1,000 kilometers adjacent, and 1,500 kilometers adjacent. The collinearity problem was also considered in our spatial autoregression models.Results： There were 175 countries included in the analysis. Of the 1,500 kilometers adjacent spatial-correlation matrix definition, removed some covariates with collinearity problem, the multivariate spatial regression analysis showed GDP (unit: per USD, estimate=0.46, p=0.033) and obesity rate (unit: per 100 persons, estimate=0.95, p<0.001) significant positive associated with T1 case rates; GDP (unit: per USD, estimate= 0.43, p=0.029) and obesity rate (unit: per 100 persons, estimate= 1.02, p<0.001) significant positive associated with T2 case rates. Of the 1,500 kilometers adjacent spatial-correlation matrix definition, removed some covariates with collinearity problem, the multivariate spatial regression analysis showed obesity rate (unit: per 100 persons, estimate=0.75, p=0.001) and female lung cancer mortality (unit: per 100 thousand, estimate=0.65, p=0.047) significant positive associated with T1 mortality; only obesity rate (unit: per 100 persons, estimate= 0.93, p<0.001) significant positive associated with T2 mortality.Conclusions： The results in our study were consistent with previous studies. After controlling for the potential confounding factors, the obesity rate was positively associated with the COVID-19 case rate and mortality. Economic or chronic disease factors were also associated with the case rate in this study. Keyword：COVID-19, pandemic, spatial autoregression models, case rate, mortality
Background： Coronavirus disease 2019 (COVID-19) was discovered in China in December 2019, and it quickly spread to the whole of China and other countries, leading to a pandemic. The first case of COVID-19 from overseas immigration occurred in Taiwan on January 21, 2020. Besides its impact on the economy, COVID-19 has also caused public panic about unknown diseases due to the pandemic in many countries and the increasing number of cases worldwide. The purpose of this study was to investigate theCOVID-19 case rate and mortality in the early stage of the pandemic and reveal associated factors by analyzing countries' data worldwide. Method： Our study used open data of countries around the world to explore the associated factors with the COVID-19 case rate and mortality, including obesity, aging, life expectancy, economic development index, literacy rate, population density, infectious diseases, and chronic diseases (the prevalence of cancer, cardiovascular disease, diabetes, tuberculosis), basic sanitation rate, medical resources (physician density, hospital bed density) were adopted by research. Spatial autoregression models were used to explore the spatial distribution of the COVID-19 case rate and mortality. The study performed spatial autoregression analysis by two index dates, the first analysis time point (July 15, 2020, T1) and the second analysis time point (December 15, 2020, T2), respectively. To explore the spatial association between explanatory variables and outcomes, spatial autoregression models by a different definition of the spatial-correlation matrix were performed, including first-order adjacent border, second-order adjacent border, 500 kilometers adjacent, 1,000 kilometers adjacent, and 1,500 kilometers adjacent. The collinearity problem was also considered in our spatial autoregression models.Results： There were 175 countries included in the analysis. Of the 1,500 kilometers adjacent spatial-correlation matrix definition, removed some covariates with collinearity problem, the multivariate spatial regression analysis showed GDP (unit: per USD, estimate=0.46, p=0.033) and obesity rate (unit: per 100 persons, estimate=0.95, p<0.001) significant positive associated with T1 case rates; GDP (unit: per USD, estimate= 0.43, p=0.029) and obesity rate (unit: per 100 persons, estimate= 1.02, p<0.001) significant positive associated with T2 case rates. Of the 1,500 kilometers adjacent spatial-correlation matrix definition, removed some covariates with collinearity problem, the multivariate spatial regression analysis showed obesity rate (unit: per 100 persons, estimate=0.75, p=0.001) and female lung cancer mortality (unit: per 100 thousand, estimate=0.65, p=0.047) significant positive associated with T1 mortality; only obesity rate (unit: per 100 persons, estimate= 0.93, p<0.001) significant positive associated with T2 mortality.Conclusions： The results in our study were consistent with previous studies. After controlling for the potential confounding factors, the obesity rate was positively associated with the COVID-19 case rate and mortality. Economic or chronic disease factors were also associated with the case rate in this study. Keyword：COVID-19, pandemic, spatial autoregression models, case rate, mortality
Description
Keywords
新型冠狀病毒, 大流行疾病, 多變數空間自迴歸分析, 確診率, 死亡率, COVID-19, pandemic, spatial autoregression models, case rate, mortality