- 學位論文
某醫學中心抗凝血藥師門診監測warfarin療效的成本效益分析
Cost-Effectiveness of the Pharmacist-assisted Warfarin Monitoring Program (PAWM) at a Medical Center
摘要
研究背景: 口服抗凝血藥品warfarin被廣泛被使用於治療或是預防血栓事件，然而其療效區間狹窄，不合適的劑量可能使病人暴露於產生嚴重出血或是血栓事件的風險之下。因此，透過個人化的治療管理才能確保warfarin的療效以及安全性，而抗凝血藥師門診監測warfarin療效 (pharmacist-assisted warfarin monitoring, PAWM) 已在許多其他國家被證實可以提升warfarin的效果及安全性並具有成本效益。 臺大醫院於2012年開辦了全國第一個病人付費之抗凝血藥師門診，而至今其療效成果及成本效益尚未與一般醫師門診照護 (usual care, UC) 進行比較。 研究目的: 了解臺大醫院抗凝血藥師門診的病人特質，並與僅接受一般醫師門診照護的病人進行比較，評估藥師抗凝血門診的療效成果及成本效益。 研究方法: 首先，以臺大醫院資料庫進行一回溯性研究，了解臺大醫院抗凝血藥師門診的病人特質、評估藥師監測warfarin的療效成果。透過此一回溯性研究所獲得的資料與其他文獻的數據，再進一步針對藥師抗凝血門診進行成本效益評估。 1.了解抗凝血藥師門診的病人特質及評估藥師監測warfarin的療效成果 收納自民國101年2月至104年4月曾接受過warfarin治療的病人，分成曾經接受過抗凝血藥師門診照護或是未接受過抗凝血藥師門診照護兩組病人。經1:4 傾向分數配對 (propensity score matching)，降低選擇性偏差，找出兩組間可以互相比較的病人。計算INR (international normalized ratio) 落於治療範圍內之時間 (time in therapeutic range, TTR)、嚴重出血事件發生率、嚴重血栓事件發生率，做為評估療效成果的依據。 2.抗凝血藥師門診監測warfarin療效的成本效益分析 本研究應用一週期長度為一個月的馬可夫模型來模擬抗凝血藥師門診監測warfarin療效相較於僅接受一般醫師照護之成本效益。本研究之馬可夫模型模擬之時間長度為20年，假設所有病人於60歲時進入研究，且在被納入研究時皆處於“無事件”的狀態，隨著時間的演變，病人可能產生嚴重血栓、嚴重出血事件、事件發生後相關的後遺症或死亡。 “無事件”狀態在本模型裡又細分成低於療效區間、 介於療效區間以及高於療效區間，以展現病人在不同的狀態下有不同的warfarin相關不良事件發生機率。抗凝血藥師門診理論上應能將病人的INR維持在療效區間的時間提高，以減少病人產生不良事件的機率。 所有套入馬可夫模型之相關數據盡量透過臺大醫院資料庫、健保資料庫得出或是引用臺灣文獻，包含事件發生機率、費用、臨床事件相關之生活品質。所有的花費計算以2012年新臺幣 (NT$) 為基準，並且引入每年3%的折現比率。 除基準值分析外，本研究亦引入單維敏感度分析(one-way sensitivity analysis)、機率性敏感度分析(probabilistic sensitivity analysis, PSA)來測試模型的不確定性是否會影響結果分析。 研究結果: 1.了解抗凝血藥師門診的病人特質及評估藥師監測warfarin的療效成果 經傾向分數配對後，共計納入1,590位病人進行結果分析，其中318個病人為接受抗凝血藥師門診照護組，1,272個病人為僅接受一般醫師照護組。以INR區間1.5-3.0而言，接受抗凝血藥師門診照護組之INR落於治療範圍內之時間明顯高於僅接受一般醫師照護組(78.4% v.s. 54.9%, p < 0.001)。即使改變INR區間至1.8-2.0 (53.3% v.s. 35.5%, p <0.001)及2.0-3.0 (33.6% v.s. 24.5%, p<0.001)，結果都顯示抗凝血藥師門診照護組之INR受到控制的情形較佳。Warfarin相關之不良事件發生率則因受限於醫院資料庫的偵測限制，一般醫師門診組的不良事件發生率可能被大幅低估，導致兩組病人之結果無法相互比較。 2.抗凝血藥師門診監測warfarin療效的成本效益分析 以健保署的觀點而言，在20年的研究區間之下，抗凝血藥師門診照護組相較於僅接受一般醫師照護組之病人平均每人可以增加0.13個品質校正生活年 (quality-adjusted life-year, QALY)及NT$53,850的額外花費，而增加成本效益比 (incremental cost-effectiveness ratio, ICER) 為NT$410,749，代表每多增加一品質校正生活年需多花費NT$410,749，小於2012年臺灣之人均國民所得 (gross domestic product per capita, GDP per capita) NT$631,142。依據世界衛生組織 (World Health Organization, WHO) 的建議，若增加成本效益比小於人均國民所得，則該醫療方案相對於其對照方案為非常符合經濟效益。 敏感度分析顯示我們的分析結果穩健度高，且以人均國民所得做為閾值的話，抗凝血藥師門診照護有86%的機率會是非常符合經濟效益。 結論: 透過臺大醫院之抗凝血藥師門診監測warfarin可以將病人的INR控制的更加良好，且相較於僅接受一般醫師門診照護來說非常符合成本效益。臺灣相較於其他歐美國家，醫療支出費用相對低廉，因此透過將INR控制良好進而減少嚴重血栓或出血事件而減少相關的花費理論上應差距不大。不過本研究結果顯示，即使在這樣的情況下再加上抗凝血藥師門診所需要增加的花費，抗凝血藥師門診仍符合經濟效益。因此，健保署應考慮給付抗凝血藥師門診監測warfarin療效的服務。
並列摘要
Background: Warfarin as a widely prescribed anticoagulant, has been proved to be effective in preventing or treating thromboembolic (TE) events. Due to the narrow therapeutic range of warfarin, inappropriate dosing may lead to major bleeding or TE events. Individualized monitoring of the international normalized ratio (INR) is required to enhance the safety and efficacy of the warfarin therapy. The pharmacist-assisted warfarin monitoring (PAWM) has been demonstrated to be effective and cost-effective in reducing warfarin-associated adverse events in Western countries. The first official pharmacist-led anticoagulation clinic offering patient-paid PAWM services in Taiwan was established in 2012 at the National Taiwan University Hospital (NTUH). To date, the clinical effectiveness and cost-effectiveness of the PAWM program compared to the usual care (UC) have not yet been evaluated. Objectives: To understand the patient characteristics, and to investigate the clinical effectiveness and cost-effectiveness of the PAWM program at the NTUH. Methods: We first used the NTUH databases to understand the features of the PAWM, including patient characteristics and clinical effectiveness. Along with these epidemiological data and data from literatures, a cost-effective analysis was performed. 1.Baseline Characteristics and Clinical Effectiveness of the PAWM A retrospective study from February 1, 2012 to April 30, 2015 was conducted in patients receiving warfarin with or without PAWM at the NTUH. A 1:4 propensity score matching was introduced to reduce the selection bias in comparing the outcomes of the two study groups. Time in therapeutic range (TTR) and major TE/bleeding even rates were captured to understand the clinical effectiveness of the PAWM compared to UC. 2.Cost-Effectiveness Analysis of the PAWM A Markov model with a one-month cycle length and 20-year time horizon was employed to assess the cost-effectiveness of the PAWM compared to the UC. Patients receiving warfarin entered the hypothetical cohort at the age of 60 and were then staying event-free or developing major TE/bleeding events followed by sequelae or death. The no-event state was split into supratherapeutic, within therapeutic, and subtherapeutic, since the patients were considered at a higher risk of encountering warfarin-associated adverse events if being out of therapeutic range. The performance of the PAWM was assumed to be in lined with the ability of keeping the patients in the therapeutic range. The likelihood of events, costs, utilities were derived from the NTUH databases, the Taiwan National Health Insurance Research Database (NHIRD) and Taiwanese literatures if applicable. All the costs were adjusted to the value of the New Taiwanese dollar (NT$) in 2012 and discounted with a rate of 3%. One-way sensitivity analysis and probabilistic analysis were performed to examine the robustness of our results. Results: 1.Baseline Characteristics and Clinical Effectiveness of the PAWM After the propensity score matching, 1,590 patients were enrolled in the study, including 318 patients in the PAWM group and 1,272 patients in the UC group. The TTR was significantly higher in the PAWM group (78.4% v.s. 54.9%, p < 0.001) when the INR range was set as 1.5-3.0. The results remained favoring the PAWM even if altering the INR range to 1.8-3.0 (53.3% v.s. 35.5%, p < 0.001) and 2.0-3.0 (33.6% v.s. 24.5%, p < 0.001). Owing to the small sample size and potential detection biases under the scope of hospital databases, the incidence rates of warfarin-associated adverse events were considered greatly underestimated in the UC group. Therefore, the incidence rates were believed to be incomparable in the two study groups. 2.Cost-Effectiveness Analysis of the PAWM From the payer’s perspective (i.e. National Health Insurance Administration, NHIA), the PAWM resulted in an average of 0.13 QALYs gained and a cost increment of NT$53,850 per person compared to UC over a 20-year horizon. The incremental cost-effectiveness ratio (ICER) was NT$410,749 per QALY gained. As the ICER was less than a gross domestic product (GDP) per capita (NT$631,142), the PAWM was considered to be very cost-effective according to the World Health Organization (WHO) recommendation. The sensitivity analyses suggested that our result was robust, and the PAWM had an 86% probability of being very-cost effective. Conclusions: The PAWM at the NTUH was proved to enhance the quality of INR control. Additionally, the PAWM was demonstrated to be very cost-effective, even if the costs saved from avoiding adverse events were thought to be minimal due to the low medical expenses in Taiwan. Our findings provide the evidence on the cost-effectiveness of the PAWM, and policy makers should consider reimbursing such service.
參考文獻
延伸閱讀
- 陳以雯（2012）。一前驅性抗凝血藥品藥師諮詢門診之成果評估〔碩士論文，國立臺灣大學〕。華藝線上圖書館。https://doi.org/10.6342/NTU.2012.11003
- 周冠妙（2006）。某醫學中心服用Warfarin之心房纖維顫動患者最低有效抗凝血強度以及維持劑量之研究〔碩士論文，國立臺灣大學〕。華藝線上圖書館。https://doi.org/10.6342/NTU.2006.02556
- Huang, H. Y. (2016). 非瓣膜性心房顫動病人使用新型抗凝血劑現況及成效分析 [master's thesis, National Taiwan University]. Airiti Library. https://doi.org/10.6342/NTU201602051
- 蘇慧真、劉盈伶、孫宗伯、周淑蓮、王致丞、蘇家震、高元、陳慧姿、廖麗怡（2018）。實證案例分享：抗凝血劑臨床停藥及使用時間之團隊決策。醫療品質雜誌，12(4)，37-39。https://www.airitilibrary.com/Article/Detail?DocID=a0000532-201807-201807260012-201807260012-37-39
- 鍾慧、楊瑛碧、高木榮、黃勝堅（2016）。Analysis of Pharmacist Interventions for Polypharmacy in the Hospital-based Accountable Family Physicians System。北市醫學雜誌，13(1)，89-94。https://doi.org/10.6200/TCMJ.2016.13.1.09