COVID-19同質及混合疫苗統計模型效益分析

Abstract

評估COVID-19疫苗效益對於防疫政策建議相當重要，然而由於多種SARS-CoV-2變異株的演化、不同時期開發的多樣化疫苗(包括阿斯特捷利康、BNT、莫德納、高端以及次世代疫苗，如：BA.1和BA.4/5等)、三劑以上的追加劑接種、接種不同基礎劑和追加劑疫苗組合的時間窗口機會(即同源和混合類型)以及從感染或COVID-19嚴重疾病中獲得的抗體和細胞介導免疫保護力衰弱，此五種與COVID-19相關的動態變化因素，導致評估COVID-19疫苗效益變得相當棘手，而此五種因素在台灣從萬華首次爆發的Alpha關注變異株開始一直到Omicron關注變異株爆發期間有所體現。為處理這五種複雜時間相關動態問題，本論文提出時間相依Cox比例風險模型進行評估。並將模型應用於台灣社區COVID-19世代追蹤數據，該數據包含個人基本資料、個人疫苗接種史以及發生COVID-19感染及其疾病進展完整訊息。本論文藉由時間相依Cox比例風險模型處理複雜且動態發展影響因素，正確評估各種疫苗組合之效益後，呈現一系列比較各種疫苗廠牌（包括同源和混合類型）對於預防感染、中/重症及死亡效益，並進一步提供不同年齡層在各種疫苗廠牌上效益進行比較。利用此評估效益結果，有助於未來新興傳染病疫苗施打政策參考。

Evaluation of COVID-19 vaccine effectiveness is critically important for informing public health policy. However, this task is unprecedentedly intractable on the grounds of five reasons. These include multiple SARS-Cov-2 variants evolution, the development of various vaccines with different timelines (including AstraZeneca, BNT, Moderna, Medigen, next generation vaccine like BA.1 and BA.4/.5 and so on), booster vaccinations with more than three doses, the time window opportunity to receive different combinations of primary and booster vaccinations, namely the homologous and the mixed type, and the waning of antibody and cell-mediated immune protection from infection or the severity of COVID-19 disease. All the five dynamic factors related to COVID-19 vaccine have been noted in Taiwan scenario from the first outbreak of Alpha Variant of Concern (VOC) in Wanhua until Omicron VOC spread over the Taiwan island. To accommodate five complex time-related dynamic issues, the present thesis proposes the time-dependent Cox proportional hazards model for evaluating the effectiveness of COVID-19 vaccine. This model is applied to a community-based COVID-19 cohort data in Taiwan, which includes comprehensive information on individuals' vaccination histories and the timing of COVID-19 infections and disease progression. Leveraging the time-dependent Cox proportional hazards model enables this thesis to accurately access the benefits of various vaccine combinations, accounting for complex and dynamically evolving influencing factors. It presents a series of results comparing the effectiveness of various vaccine brands (including homologous and mixed types) in preventing infection, moderate to severe disease, and mortality. Furthermore, the thesis provides comparisons of vaccine effectiveness across different age groups. The results of this evaluation offer new insights into future vaccination policies for emerging infectious diseases.