药物经济学概率敏感性分析中关联参数的处理与实践探讨

Discussion and Practice of Correlating Parameters in Sensitivity Analysis of Pharmacoeconomics

在基于模型法的药物经济学评价中,需要通过设定参数的假定分布进行概率敏感性分析以验证分析结果的稳健性。若模型参数间存在一定的相关性,则需要对关联参数进行处理以保证概率敏感性分析结果的准确。本研究从理论出发,论述了在药物经济学评价中进行关联参数校正的必要性,并在此基础上结合实际案例操作,提供一套较为完整的如何从生存曲线得到经校正的转移概率,并提供相应的计算公式。由Stata获取参数的方差-协方差矩阵,使用Excel对回归参数使用Cholesky分解进行校正处理,最终得到Markov模型在概率敏感性分析中的转移概率,以期为开展更为规范的药物经济学评价提供参考。

In the model-based cost-effectiveness analysis,it′s necessary to do the probabilistic sensitivity analysis by setting the assumed distribution of related parameters in order to verify the robustness of the model results.If there is a certain correlation between parameters,those parameters are needed to be corrected to make sure the accuracy of the results of the probabilistic sensitivity analysis.Based on the theory,this study discusses the necessity of correlating parameters correction in the economic evaluation combined with the actual case analysis,providing a relatively complete process of how to obtain the corrected transition probability from the survival curve.Also,this article provides the relevant calculation formula to help reader understand and practice easier.The variance-covariance matrix of the parameters is obtained by Stata,and the regression parameters are corrected by Cholesky decomposition using Microsoft Excel.Finally,the transition probability of the Markov model in the probability sensitivity analysis is obtained.This article will help to improve guidance and support in economic evaluation and develop a more standardized sensitivity analysis process for researchers in the future.