| 摘要: | 研究背景： 在2010年世界衛生組織的研究中，子宮頸癌在婦女癌症中的排名高居所有癌症中的第三名，並且在2010年全世界因子宮頸癌死亡人數高達二五萬人以上。研究顯示，施打人類乳突病毒疫苗可預防人類乳突病毒所引起的子宮頸癌。目前有很多篇關於人類乳突病毒的經濟評估模型研究，卻缺少系統性文獻回顧的文獻。研究者通常不清楚要用哪些模型來執行關於人類乳突病毒疫苗的經濟評估研究，所以有一篇對於人類乳突病毒研究的系統性文獻回顧或許可以提供決策者所需要的模型建構基本知識。
研究目的： 提供給未來台灣的研究者要做此類關於人類乳突病毒疫苗施打與子宮頸癌篩檢併行策略之經濟效益評估需要的模型類型與基本訊息。
研究方法： 本研究蒐集西元2003年1月到2012年12月所有關於在女性的子宮頸癌篩檢策略中，與人類乳突疫苗施與子宮頸癌篩檢策略併行做為防制子宮頸癌的經濟評估報告，並搜尋The Cochrane Library, NHS Economic Evaluation Database (NHS EED), The CEA Registry和PubMed等四大學術資料庫，符合搜尋條件的共有67篇(n=1000)。在所有結果中有21篇使用動態模型，45篇使用馬可夫模型，1篇兩種模型都有使用。在所有的研究中針對人類乳突病毒疫苗和子宮頸篩檢策略併行與子宮頸癌篩檢比較的所有相關假設與結果製作成簡易表格，藉以分析出使用模型與假設和結果之關係。
研究結果： 本研究發現動態模型與馬可夫模型為兩種主要用來分析人類乳突病毒疫苗的模型；馬可夫模型主要用來分析一個世代並具有可解釋性的優點，動態模型則是用來分析多世代和分析群體免疫主要會使用的模型，並發現群體免疫在特定的假設方法下也可以用馬可夫模型來評估。在結果方面，幾乎所有的研究都顯示，如果以人類乳突病毒疫苗與子宮頸癌篩檢策略併行和只有子宮頸癌篩檢策略做比較是具有經濟效益的。在所有67篇文獻的敏感度分析方面，折現率參數和疫苗的價錢參數對於成本效果比的影響是最顯著的。所有的研究顯示雖然所有研究所使用的模型、假設、資料來源等都不盡相同，人類乳突病毒疫苗施打與子宮頸癌篩檢併行是對於子宮頸癌預防具有經濟效益潛力的介入方法。
研究結論： 本研究顯示馬可夫模型和動態模型是研究人類乳突病毒疫苗施打與子宮頸癌篩檢併行策略之經濟效益評估的主要方法，然而並不限定於分析群體免疫一定要使用動態模型，在馬可夫模型中也可透過其它假設來分析群體免疫所造成的影響。
Background: Cervical cancer is the fourth leading cause of female cancer deaths in the world, which accounts for 274,883 deaths in 2010 and it also ranks the third most common cancer in women. A Human Papillomavirus vaccine has been used to prevent HPV induced cervical cancer in most countries. More and more cost-effectiveness studies on a human papillomavirus vaccine are undertaken around different countries and regions. A formal up-to-date systematic review can be helpful for future researchers in Taiwan.
Objective: To provide information about methods and fittest model choices to future researchers if they want to conduct a cost-effectiveness analysis of HPV vaccination adding to screening programs comparing with screening alone in Taiwan.
Search methods: The Cochrane Library, NHS Economic Evaluation Database (NHS EED), The CEA Registry, and PubMed were searched.
Selection criteria: The cost-effectiveness analysis of human papillomavirus vaccination in females.
Data collection and analysis: The full-text of articles were extracted. Data and outcomes were collected. Economic analyses were summarized in narratives and table forms. Biases were being assessed using a Phillips checklist.
Main results: 67 published studies in the period of 2003 January to 2012 December were eligible for this review. The studies all use mathematical models. Among them, 21 studies use dynamic models, 45 studies use static Markov models, and one study use both a dynamic and a Markov model. The dynamic models and the Markov models are two main types of analyzing the cost-effective of adding a HPV vaccination to screening programs. A Markov model is suitable for studies of a single cohort and if has the advantage of transparency. A dynamic model is suitable for multi-cohort studies and for analyzing herd immunity effects, but there are other methods for the Markov models to access for herd immunity. Despite the different methods and different places where studies took place, the base-case results of all studies indicating that incorporating HPV vaccination to screening programs might be cost-effective. Subgroup analyses are conducted to further understand the impact of different tools, populations, and underlying screening programs. In the sensitivity analysis results of almost all studies, discount rates and vaccine prices are two major parameters sensitive to incremental cost-effectiveness ratios. Among all 67 studies, except for six studies in the Netherlands, Thailand, Taiwan, US, and Australia, it is generally cost-effective to add a Human papillomavirus vaccine into screening practices in many countries according to the this review.
Conclusion: For analyzing the cost-effectiveness analysis of HPV vaccination adding to screening programs, a Markov model and a dynamic model are two major model types to simulate the nature history of cervical cancer, but it is not necessary to use dynamic models to assess herd immunity effects; Markov models are capable of analyzing herd immunity through other assumptions. |
