【摘要】：目的旨在探討當(dāng)進(jìn)行藥物經(jīng)濟(jì)學(xué)評(píng)價(jià)時(shí),如何在敏感度分析中選擇合適的參數(shù)分布。方法針對(duì)各種類型參數(shù),根據(jù)其限定條件,結(jié)合分布特征,選擇合適的參數(shù)分布假設(shè)。結(jié)果概率參數(shù)只能在0到1之間取值,互斥時(shí)間概率的總和應(yīng)等于1;相對(duì)風(fēng)險(xiǎn)參數(shù)的可信區(qū)間是通過對(duì)數(shù)轉(zhuǎn)換所得;成本參數(shù)取值范圍為0至無窮大;效用參數(shù)通常取值0至1(特殊情況低于0)。結(jié)論概率參數(shù)可假設(shè)為Beta分布,相對(duì)風(fēng)險(xiǎn)參數(shù)可假設(shè)為對(duì)數(shù)正態(tài)分布,成本參數(shù)可假設(shè)為伽馬分布或?qū)?shù)正態(tài)分布,效用參數(shù)可假設(shè)為Beta分布、對(duì)數(shù)正態(tài)分布或伽馬分布。
[Abstract]:Objective to explore how to select appropriate parameter distribution in sensitivity analysis when evaluating pharmacoeconomics. Methods according to the limited conditions and the distribution characteristics, a suitable parameter distribution hypothesis was selected for all kinds of parameters. Results the probability parameters can only be taken from 0 to 1, and the sum of probability of mutual exclusion time should be equal to 1. The confidence interval of relative risk parameters is obtained by logarithmic transformation, and the range of cost parameters is from 0 to infinity. Utility parameters usually range from 0 to 1 (less than 0 in special cases). Conclusion the probability parameter can be assumed as Beta distribution, the relative risk parameter can be assumed to be lognormal distribution, the cost parameter can be assumed as gamma distribution or lognormal distribution, and the utility parameter can be assumed to be Beta distribution, lognormal distribution or gamma distribution.
【作者單位】： 中南大學(xué)湘雅二醫(yī)院藥學(xué)部;中南大學(xué)湘雅二醫(yī)院PET-CT中心;
【基金】：國家自然科學(xué)基金(81173028) 青年科學(xué)基金項(xiàng)目(81401547)
【分類號(hào)】：R95
，

本文編號(hào)：2457193