請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/69612
標題: | 流感病毒株同期競爭現象與當代流感疫苗政策之數理模建 Mathematical Modelling for Co-circulating Influenza Strains Competition and Contemporary Influenza Vaccination Policy |
作者: | Bin-Shenq Ho 何秉聖 |
指導教授: | 趙坤茂(Kun-Mao Chao) |
關鍵字: | 流感,傳播,病毒株競爭,疫苗接種政策,數理模建, influenza,transmission,strain competition,vaccination policy,mathematical modelling, |
出版年 : | 2018 |
學位: | 博士 |
摘要: | 傳染病傳播動力學,乃傳染病防治的基本課題之一,支配同期流傳的流感病毒株在人類社會中的興衰變遷。流感防治政策有關健康和經濟的影響評估,係嵌植於傳染病傳播動力學的基本架構之中進行演算推估。二十一世紀初以來,為了減輕大流行性流感可能造成的破壞,其相關之數理模建,已成為大流行性流感防範整備各領域之中,較具研究成果的領域之一。 數理模建發展的貢獻,不僅可洞察流感防治未見之處,更提供流感大流行與季節流行未明謎團闡明之道。
多種流感病毒株同期流傳,是自然界中的常態;何種病毒株可勝出,將決定季節性流感甚至大流行性流感的發展趨勢。同期流傳的流感病毒株之競爭現象,以往雖然認為與該等病毒株之興衰變遷密切關連,然而,此一現象鮮少進行人類族群層次的量化研究。另一方面,作為流感防治基石的疫苗接種政策,總是面臨病毒株不斷變化的不確定性,以及從而產生疫苗不相匹配的可能性,導致當代流感疫苗接種政策的成效,其看似矛盾的評估資訊,經常阻礙了公眾進行充分知情風險溝通的機會。 為了探討上述兩個議題,我們建構異構動態傳播模型,並以整個流感季每週的A/H1N1流感病毒分離率進行模型擬合。模型建構過程之中,完成我國2007-2008年單支群流感季實境模擬測試之後,再加入2008-2009年A/H1N1流感病毒各單一支群流行曲線進行擬合過程,以建構2008-2009年雙支群流感季實境模擬,並據以量化方式探索解讀同期流傳流感病毒株的競爭動態。我們進一步充分利用異構動態傳播模型,運用其年齡特異化傳播結構,採用全國流感實驗室數據暨監測系統資訊,重建我國2007-2008年流感季實境,作為疫苗接種情境各項方案比較的基準。有關當代流感疫苗接種政策的評估,本研究著重於其對於季節性流感的影響,並在整體人口族群層次,進行有關流行高峰減壓、疫情負擔輕減、高峰時間變化等三項疫病流行度量的影響分析。 研究結果顯示,異構動態傳播模型可發掘潛在最佳的擬合情境,其與流感季實境間之相關係數可高達96%,而且,所有成功的模擬情境均收斂到最佳擬合情境,異構動態傳播模型並可估算同期流傳流感病毒株的個別年均有效傳染數。2008-2009年流感季期間,後繼出現主宰季節流行的A/H1N1流感病毒2B-2支群,其神經胺酸酶攜帶H275Y突變,年均有效傳染數估計約1.65,至於先行出現的A/H1N1流感病毒2C-2支群,其年均有效傳染數起初相當於1.65,但在2B-2支群浮現之後,2C-2支群被取而代之,其年均有效傳染數最終呈現約為0.75。異構動態傳播模型顯示,2B-2支群勝出於2008-2009年流感季,主要係因2C-2支群遭遇2B-2支群之際,其傳播適能降低了大約71%之多。 進一步研究結果證實,流感疫苗量能的建置雖然不可或缺,但卻無從擔保可對季節性流感疫情發揮實質而顯著的影響,倘若發生疫苗與盛行的病毒株不相匹配的情形,即可能大幅抵銷建置疫苗量能的效益。儘管如此,經由數理模建論證,疫苗前置接種可以彌償些許疫苗效能表現不佳的窘境,此外,如果疫苗與盛行的病毒株匹配良好,疫苗前置接種更可增益其成效。至於疫苗分配優化問題,我們的研究結果發現,流感傳播動力取決於各年齡群內與各年齡群間的傳播綜效,就減輕整體人口族群流感傳播的考量而論,優先接種對象並非6至12歲年齡群,而是20至39歲年齡群暨13至19歲年齡群。 總結而言,我們的研究運用跨科際數理模建工程,利用數據驅動提升問題解析能力,揭露了攜帶H275Y突變的A/H1N1流感病毒株,其於2007-2009年流感季期間,竄流全球詭祕的傳播動態。這將啟發我們處理2009年流感大流行之後,季節持續流行的A/H1N1pdm09流感病毒株,其抗藥性病毒株不斷浮現的議題。研究顯示,我們通過數理模建提供了一種前瞻性的方法,解決在人類族群層次上看似無法理解的問題,也期望彌合相關學門的解析能力,應用於生物分子層次上的議題。再者,考慮疫苗量能建置以及潛在疫苗不匹配等問題,疫苗前置接種應列為推動季節性流感防治政策的重要工作,而數理模建可以提供政策制定者一個實質的平臺,進行具體的衛生政策影響評估,並據以加強與社會公眾之間的風險溝通。 Understanding transmission dynamics is one of the fundamental problems in infectious diseases control. Transmission dynamics governs the vicissitudes of co-circulating influenza strains in the human population and constitutes the essential framework that embeds the health and economic impact of influenza control policy. With an eye to mitigating the devastation of pandemic influenza, mathematical modelling pertinent to influenza pandemic has been one of the more productive fields in the pandemic preparedness since the early 21 century. The contribution not only provided insights into influenza control but also shed light on the resolution of some enigmas of influenza pandemic and epidemic. Co-circulation of influenza strains is the rule. Which strain could outcompete would determine the trend of the seasonal epidemic and even the worldwide pandemic. Although competition was thought to be involved in the vicissitudes of co-circulating influenza strains, the phenomenon was rarely studied quantitatively at the human population level. On the other hand, vaccination policy, being the cornerstone of influenza control, always faces the uncertainty of the evolving virus and consequently the vaccine mismatch that probably ensues. As reported favorably or unfavorably, the seemingly paradoxical impacts of the contemporary influenza vaccination policy usually hindered the public from risk communication under well informed situation. To deal with the problems stated above, we constructed a heterogeneous dynamic transmission model and ran the model to fit the weekly A/H1N1 influenza virus isolation rate through an influenza season. The construction process started on the 2007-2008 single-clade influenza season and, with the contribution from the clade-based A/H1N1 epidemiological curves, advanced to the 2008-2009 two-clade influenza season. The main purpose was to decipher the competition dynamics of co-circulating influenza strains in a quantitative way. Furthermore, to assess the impact of the contemporary influenza vaccination policy, we exploited the age-structured dynamic transmission model by using the laboratory data of the national influenza surveillance system to reconstruct the 2007-2008 baseline scenario with which the vaccination scenario of interest could be compared. The impact on the seasonal epidemic was analyzed in terms of peak decompression, burden reduction, and peak timing change at the population level. Our model found the potentially best-fit simulation with correlation coefficient up to 96% and with all the successful simulations converging to it. The annual effective reproductive number of each co-circulating influenza strain was estimated. During the 2008-2009 influenza season, the annual effective reproductive number of the succeeding A/H1N1 clade 2B-2, carrying H275Y mutation in the neuraminidase, was estimated around 1.65. As to the preceding A/H1N1 clade 2C-2, the annual effective reproductive number would originally be equivalent to 1.65 but finally took on around 0.75 after the emergence of clade 2B-2. The model reported that clade 2B-2 outcompeted in the 2008-2009 influenza season mainly because clade 2C-2 suffered a reduction of transmission fitness of around 71% on encountering the former. Further study confirmed that vaccine capacity building, though indispensable, could not guarantee substantial impact on the seasonal influenza epidemic. Vaccine mismatch might greatly offset vaccine capacity building. Nevertheless, our model demonstrated that advance vaccine distribution could at least compensate for some vaccine underperformance. Besides, in the case of a well-matched vaccine, advance vaccine distribution could even potentiate its performance. As to vaccine allocation optimization problem, we found that within-group contact rate and the between-group contact rate operated on transmission dynamics jointly. In terms of transmission mitigation, our study indicated that the age group of 20-39 years as well as 13-19 years was the priority for allocating vaccines, rather than the age group of 6-12 years. We conclude that interdisciplinary data-driven mathematical modelling could bring to light the transmission dynamics of the mysterious A/H1N1 H275Y strains during the 2007-2009 influenza seasons worldwide and may inspire us to tackle the continually emerging drug-resistant A/H1N1pdm09 strains. As a matter of fact, we provide a prospective approach through mathematical modelling to solving a seemingly unintelligible problem at the human population level and look forward to its application at molecular level through bridging the resolution capacities of related disciplines. Furthermore, our research would suggest that, taking vaccine capacity building along with potential vaccine mismatch into consideration, advance vaccine distribution be put high on the agenda of seasonal influenza control policy. By way of mathematical modelling, we provide a tangible platform for policy makers to evaluate health policy impact and to enhance risk communication with the public. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/69612 |
DOI: | 10.6342/NTU201801011 |
全文授權: | 有償授權 |
顯示於系所單位: | 資訊工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-107-1.pdf 目前未授權公開取用 | 1.99 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。