基于比例优势模型和优化试验设计的计算机临床

【摘要】 基于比例优势模型的临床试验仿真，给出优化试验设计方法，以抗偏头痛新药那拉曲坦(naratriptan)为例，仿真药物临床试验结果。采用比例优势模型，分析不同给药时间、剂量对疼痛减轻(pain relief，PR)的影响。然后，采用优化实验设计最大化FIM行列式，比较临床试验仿真的设计结果，得出较合理的给药剂量和时间的设计方案。通过曲线图直观地表示出PR概率的变化趋势，反映了给药时间和给药剂量对PR产生的影响。从仿真结果得出，药物效应的变化滞后于血药浓度的变化，疼痛减轻概率和给药剂量、采样时间、模型参数等因素有关。临床实验仿真对于新药的研发是一种有效的工具，它可以定量估计可控因素对治疗效果的影响。

【关键词】 比例优势模型;D-优化试验设计;Fisher信息矩阵;计算机临床试验仿真

　Clinical Trial Simulation based on

　　Proportional Odds Model and D瞣ptimal Design ZHENG Dan1, ZHU Ling, LIU Yajie, SHI Xinling

　　(1.Electronical Engineering Department, Information School, Yunnan University, Kunming 650091，China)

　　Abstract：To introduce the clinical trial simulation based on proportional odds model. Taking naratriptan as example, which is a novel agonist for the acute treatment of migraine, we adopted the method of D-optimal design to simulate the drug clinical trail of naratriptan.We adopted proportional odds model and analysed different factors such as time and does how to impact Pain Relief(PR). Then, the D-optimal design was used to maximize the determinant of Fisher information matrix(FIM) and compared the clinical trial simulated results. Finally more reasonable dose and time of the designs were obtained.The curve of the PR probability expressed PR trend directly, reflecting administration dose and time on the impact of PR. According to the simulated results, we found that there was a delay of the effect with respect to the plasma concentration in “effect” compartment and the probability of having a specific Pain Relief (PR) score depended on different facts such as dose, the effect sampling time and model parameters. Clinical trial simulation is a useful tool for the new drug development and it can be used for the quantitative assessment of the controllable factors to the effect of treatment.

　　Key words：Proportional odds model; D-optimal design; Fisher information matrix; Clinical trials simulation

　　1 引 言

　　计算机临床实验仿真(CTS)运用计算机模拟技术从前期研究中获得信息，揭示试验设计中变量和假设对结果的影响，预测和评价不同研究方案可能产生的结果。CTS适用于实际系统费用昂贵，存在安全问题的情况，根据药物的药效学和药代动力学设计临床试验仿真，分析仿真结果，改进方案，从而节约了成本，提高了研发新药的效率[1-2]。

　　在临床试验中常遇到反应变量为多分类有序变量[3]，采用Logistic回归模型研究多分类反应变量与其影响因子间关系，分析得到调整后的药物评价结果[4]。D-优化试验设计是按照一定的算法，最大化FIM行列式，在几组设计方案中寻求最优。本研究提出了将传统的比例优势模型与优化试验设计结合的方法，进行计算机药物临床试验仿真。

　　2 方法