POPULATION PHARMACOKINETIC MODELING ON A SUPERCOMPUTER

描述（取自项目摘要）：在此资源相关的 研究项目，我们建议：1）在超级计算机上实现我们的 当前人口的非参数EM和迭代贝叶斯软件 药代动力学建模，目前在PC上运行; 2）实现扩展 该软件包括大型非线性药代动力学和 药物作用和毒性的药效学（PK/PD）模型; 3）开发和 采用新软件计算最佳信息时间表， 监测血清药物水平和测量药物作用的方案， 多中心临床药物试验;和4）开发和实施友好的 图形用户界面（GUI）和通信包，用于访问 超级计算机在万维网和互联网上，作为一个快速和 有效的研究资源。 群体PK/PD建模是理解药物在体内如何发挥作用的关键。 人 它提供了基本的结构， 临床药物试验，以及临床药物试验的智能设计 来评估疾病的治疗过程。 但目前 这种建模工作中的两个重要瓶颈：1）当前的软件 对于研究人员来说使用起来并不友好，2）需要很长时间才能 分析大型数据集。 由于程序通常在PC上运行， 研究人员在数据集的数量上受到了很大的限制， 他们可以分析的大型临床试验。 一个国家级的研究资源，在研究人员的PC上有一个友好的GUI， 可以很容易地连接到一个超级计算机，可通过网络和 互联网，将是非常有用的许多工作人员在国家卫生研究院，在学术 中心和制药行业。 有了这些资源， 能够轻松快速地分析数据并估计人口 参数分布和单个参数值。 此外，该软件提出，以获得最翔实的， 用于测量血清水平和药物作用的成本效益策略， 治疗和毒性，将提供一种简单和快速的方法， 最具成本效益的多中心药物试验方案设计， 浓度控制的临床试验。 人口模型和最佳监测时间表也将很容易 与现有软件的链接，用于新的“多模型”药物方案设计 为了最佳精确地实现期望的临床目标， 患者变异性，包括“浓度控制”和“效应” “临床药物试验”。

DESCRIPTION (Taken from project abstract): In this resource-related research project, we propose to: 1) Implement on a supercomputer our current nonparametric EM and iterative Bayesian software for population pharmacokinetic modeling, currently running on PC's; 2) Implement extensions of this software to include large nonlinear pharmacokinetic and pharmacodynamic (PK/PD) models of drug effect and toxicity; 3) Develop and implement new software to compute optimally informative schedules and protocols for monitoring serum drug levels and measuring drug effects in multicenter clinical drug trials; and 4) Develop and implement a friendly graphical user interface (GUI) and communication package for access to the supercomputer over the World Wide Web and the Internet, as a fast and efficient research resource. Population PK/PD modeling is the key to understanding how drugs behave in people. It provides the basic structure for understanding what happens in clinical drug trials, and for the intelligent design of clinical drug trials to evaluate treatment of a disease process. However, there are currently two significant bottlenecks in such modeling efforts: 1) current software has been unfriendly for researchers to use, and 2) it takes a long time to analyze large data sets. As the programs are usually run on PC's, investigators have been significantly limited in the number of data sets from large clinical trials which they can analyze. A national research resource with a friendly GUI on the researcher's PC that can easily be connected to a supercomputer, available over the Web and the Internet, will be highly useful to many workers in the NIH, in academic centers, and in the pharmaceutical industry. With this resource they will be able easily and rapidly to analyze their data and estimate population parameter distributions and individual parameter values. Further, the software proposed to obtain the most informative and cost-effective strategies for measuring serum levels and drug effects, both therapeutic and toxic, will provide an easy and rapid method for optimal and most cost-effective design of protocols for multicenter drug trials and concentration controlled clinical trials. The population models and optimal monitoring schedules will also readily link to existing software for new "multiple model" design of drug regimens for optimally precise achievement of desired clinical goals, minimizing patient variability, both in "concentration controlled" and in "effect controlled" clinical drug trials.