Aarons L
School of Pharmacy and Pharmaceutical Sciences, University of Manchester, England.
Clin Pharmacokinet. 1999 Apr;36(4):255-64. doi: 10.2165/00003088-199936040-00001.
Pharmacokinetic-pharmacodynamic modelling is being used increasingly as a tool in drug development because often in phase III clinical trials only sparse data are available for analysis and so a nonlinear mixed effects modelling approach has to be adopted. Specialist data analytical techniques and software are required to analyse such data. This article reviews some of the software currently available for performing nonlinear mixed effects modelling. A questionnaire was devised and sent to a number of software producers and the findings are presented and discussed in this paper. The programs could be grouped into 3 main categories: parametric and nonparametric maximum likelihood and Bayesian. It was apparent from the questionnaire that software development for population data analysis is a very active area of investigation. The implementation of methodologies varied widely between the packages: some were self-contained programs, whereas others were written within another application, usually a statistical package. They also varied with respect to their ease of use and level of support offered by the software producers. Although robustness and reliability are important concerns, they were not addressed in the present review. Most of the programs surveyed are in continual development.
药代动力学-药效学建模正越来越多地被用作药物研发的工具，因为在III期临床试验中，通常只能获得稀疏的数据用于分析，所以必须采用非线性混合效应建模方法。分析此类数据需要专业的数据分析技术和软件。本文综述了目前可用于进行非线性混合效应建模的一些软件。设计了一份问卷并发送给多家软件生产商，本文将呈现并讨论调查结果。这些程序可分为三大类：参数和非参数最大似然法以及贝叶斯法。从问卷中可以明显看出，群体数据分析的软件开发是一个非常活跃的研究领域。各软件包之间方法的实现差异很大：有些是独立的程序，而有些则是在另一个应用程序（通常是统计软件包）中编写的。它们在易用性和软件生产商提供的支持水平方面也有所不同。尽管稳健性和可靠性是重要的关注点，但在本综述中并未涉及。所调查的大多数程序都在不断发展。
