根據聯結邊際廣義伽瑪分配的關聯結構函數研究成本效益比率; A study of cost-effectiveness ratio based on Copula functions with generalized gamma marginal distributions

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/48215

Title:	根據聯結邊際廣義伽瑪分配的關聯結構函數研究成本效益比率;A study of cost-effectiveness ratio based on Copula functions with generalized gamma marginal distributions
Authors:	張稚辰;Chih-chen Chang
Contributors:	統計研究所
Keywords:	成本效益增量比率;淨健康利益增量;關聯結構函數;廣義伽瑪分配;成本效益分析;成本效益平面;cost-effectiveness plane;ICER;INHB;generalized gamma distribution;Copula functions;CEA
Date:	2011-06-23
Issue Date:	2012-01-05 14:41:39 (UTC+8)
Abstract:	本文回顧近年來針對醫療成本效益分析的研究，比較傳統上所使用的方法如成本效益平面（cost-effectiveness plane）、成本效益增量比率（incremental cost-effectiveness ratio,記做ICER）、淨健康利益增量（incremental net health benefit,記做INHB）的優劣及特色。然而本文認為傳統方法具有不易解釋以及無法同時比較多個治療方案成本效益的困擾，因此提出單一治療方案成本效益比率（ratio of cost-effectiveness ratio,記做RCE）的概念做為新的評量準則。本研究中收集的資料為每一病患的存活時間和接受治療所付出的醫療成本，是為成對資料。由於在實務上，成本及存活時間大多為右偏分布，在考量兩者間具有相關性的情況下，提出使用適當的關聯結構函數建構二元廣義伽瑪分配。然後在此架構下利用最大概似法估計RCE且提供建構RCE信賴區間方法。本文在不同的關聯結構函數、相關係數、設限比例等條件下，藉由模擬研究探討本文所提供建構信賴區間方法的覆蓋機率、長度以及信賴上、下界的錯誤覆蓋率表現，並於最後提出一例子作為說明。 Reviewing researches in last decades on medical cost-effectiveness analysis (CEA), we analyze traditional measurements, e.g. cost-effectiveness plane, incremental cost-effectiveness ratio(ICER), and incremental net health benefit(INHB). These measurements are difficult to interpret and to apply for comparing multiple diagnostic methods. Therefore, we suggest the ratio of cost-effectiveness(RCE) as a new criteria. Which is the ratio of the cost of taking a certain medical therapy for a patient and the patients’ survival time. In practical cases, the distributions of the correlated cost and survival time are generally right-skewed. Therefore, we employ appropriate copula to construct the joint generalized gamma distribution. Under the joint distribution, we find the maximum likelihood method estimate and hence the confidence interval for the RCE. The results of a simulation investigation of the coverage probability, interval length, lower and upper error rate of confidence interval for different censoring probabilities and degrees of correlation in several possible copulas functions are reported. Finally, the proposed method is illustrated by using an example.
Appears in Collections:	[Graduate Institute of Statistics] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	778	View/Open

社群 sharing

Loading...