應用傳染病數理模型SEIR模擬冰島以檢驗取代檢疫之邊境管理 對台灣之本土感染COVID-19病例和人群擴散之影響

Abstract

背景及目的 COVID-19全球疫情持續惡化，各國政府控制疫情仍然仰賴各種防疫管制措施，其中相較國際疫情輕微的我國以邊境管理制最為重要，但是為期14天的檢疫隔離會造成個人負擔，亦影響社會和經濟。部分學者，認為應以入境全面篩檢取代14天居家檢疫。本研究應用傳染病數理模型SEIR模擬SARS Co-V2來自境外移入病例和人群擴散的情況，以此估算在採用冰島入境檢疫方式，可能對於本土傳播之影響。 方法 假設每日1500人入境，其中有10名COVID-19感染者入境台灣，於入境時接受COVID-19檢驗，檢驗結果陽性者須接受隔離，結果陰性者可自主健康管理14日，不須接受居家檢疫。入境結果陰性者可能陸續進入可傳染期的症狀感染前期，部分個案之後出現症狀，而後經由公衛人員追蹤將安排就醫隔離，部分個案未就醫則未被診斷和隔離。有傳染力的感染者進入社區後與社區之易感受者接觸後發生疫情擴散。以此模型模擬45天後，將累積多少新增本土確診個案。基於這些假設和簡化，應用isee system之STELLA Professional Version 1.9.5. 軟體構建一個COVID-19的SEIR模型框架。參數設定除參考文獻外，我們也使用截至2020年7月29日之台灣確診個案資料(380名境外移入個案)，分析彙整個案之入境日、發病日或檢驗日，估計入境當天不同疾病階段的分佈比例，以設定參數。 結果 第一部分模擬不同R0估計本土個案數，R0為影響模型之關鍵，R0越大本土疫情愈嚴重；第二部分模擬有症狀者就醫比率及接觸者追蹤達成率不同之影響，即使在兩者比率皆100%狀況下，在R0為2時，估計仍有高達83名本土個案。第三部分模擬不同無症狀感染者比例和無症狀感染者相對傳染性之影響效應，發現有症狀個案比例越高，估計本土個案數越多；無症狀比例低且無症狀傳染力高者之情境估計之個案數多於無症狀比例高而傳染力低之情境。 結論: 以檢驗取代檢疫可能造成防疫破口而引起本土疫情，建議仍須採用檢疫搭配症狀監測或檢疫期滿檢驗以降低傳播風險。無症狀個案在傳染力增加時對傳播有關鍵影響，未來需考慮增加相關措施，以補足症狀監測之不足。

Background and purpose The global epidemic of COVID-19 continues to worsen, and governments of various countries still rely on various epidemic preventive measures to control the epidemic. Among them, the border management system is the most important measures for our country. However, the 14 days of quarantine will cause personal burdens and affect the society and the economy. Some scholars believe that the 14 days of quarantine should be replaced by a comprehensive entry screening. In this study, the SEIR mathematical model of infectious diseases was used to simulate if we adopt the entry quarantine method in Iceland to estimate the possible impact on local transmission f SARS Co-V2 from the imported cases. Methods: Assuming that 1,500 people enter Taiwan every day, 10 of them are infected with COVID-19 entering Taiwan and undergo COVID-19 testing upon entry. Those with a positive test result must be quarantined, and those with a negative result can manage their own health for 14 days without quarantine. Those with negative entry results may gradually enter the infectious pre-infection stage. Some cases develop symptoms afterwards, and then follow up by public health personnel to arrange medical treatment and isolation, and some cases have not been diagnosed and isolated. The epidemic spreads after the infectious ones enters the community and contact the susceptible in the community. Using this model to simulate how many new local cases will accumulate after 45 days. Based on these assumptions and simplifications, the STELLA Professional Version 1.9.5. isee system is used to build a SEIR model framework for COVID-19. In addition to the parameters adopted from literature review, we used the data about confirmed cases in Taiwan to set parameters. As of July 29, 2020, there were 380 imported cases. We analyzed the entry date, onset date or testing date of the cases to estimate the distribution ratio of different disease stages on the day of entry. Result The first part is to simulate different R0 to estimate the number of local cases. R0 is the key to the model. The local epidemic is more serious when the R0 increase. The second part is to simulate the impact of the different rates of symptomatic patients seek medical help and the contact tracing attainment rates. Even both parameters reached 100%, if the R0 is 2, it is estimated that there are still as many as 83 local cases. The third part is to simulate the effect of the proportion of asymptomatic infections and the relative infectivity of asymptomatic infections. It is found that the higher the proportion of symptomatic cases, the more local cases are estimated. The estimation showed more cases in the situational with low asymptomatic proportion and high asymptomatic infectivity than those with a high proportion of asymptomatic and a low infectivity. Conclusions: Replacing quarantine with screening may cause a breach in epidemic prevention and cause a local epidemic. It is recommended that quarantine together with symptom monitoring or COVID-19 testing after quarantine to reduce the risk of SARS Co-V2 transmission. Asymptomatic cases have a critical impact on transmission when the infectivity increases. In the future, it is necessary to consider adding relevant measures to supplement the lack of symptom monitoring.