基于遗传算法的威布尔溶出曲线参数估算方法研究

目的:利用遗传算法全局搜索能力的优势,进行威布尔溶出曲线参数估算。方法:以通脉丸中竹节香附素A的溶出数据为研究载体,利用Matlab软件中遗传算法优化工具箱实现威布尔溶出曲线参数估算,并与非线性最小二乘法估算结果进行统计学比较。结果:遗传算法与非线性最小二乘法估算的威布尔溶出曲线参数无显著性差异(P>0.05),而遗传算法估算参数的校正均方差小于非线性最小二乘法。结论:遗传算法适用于威布尔溶出曲线参数的估算。     

基于1stOpt软件的威布尔溶出曲线拟合方法研究

目的:利用1stOpt软件实现威布尔溶出曲线拟合,提供一种新的可获得溶出曲线参数的方法。方法:以通脉丸中芍药苷的溶出数据作为研究对象,利用1stOpt软件中麦夸特法+通用全局优化(LM-UGO)法来实现威布尔溶出曲线拟合,并与Matlab软件中的非线性最小二乘(NLLS)法及遗传算法拟合结果进行比较。结果:LM-UGO、NLLS和遗传算法拟合得到的威布尔溶出曲线参数无显著性差异(P>0.05),而溶出参数的校正均方差分别为0.00990、0.01212、0.01003,以LM-UGO法值最小。结论:1stOpt软件具有操作简单、拟合快速而准确、收敛速度快等优点,是威布尔溶出曲线拟合及参数估算的有力工具。     

一元加权最小二乘估计处理异方差性及SAS实现

分析线性回归分析中异方差性的产生原因和影响,介绍加权最小二乘估计法处理一元回归分析中异方差性的原理,并用SAS软件加以实现。通过比较SAS软件输出结果,得出当存在异方差性时,加权最小二乘估计法比普通最小二乘估计法估计效果更好的结论。     

自回归条件异方差模型在医学数据分析中的应用

为探讨自回归条件异方差模型在医学数据分析中的应用效果 ,采用自回归条件异方差模型对实例进行了分析 ,并与多元回归分析结果进行了比较。结果显示 ,当数据方差不齐时 ,自回归条件异方差模型能够较好的应用于该类资料的分析     

最小二乘法在医院药学管理中的应用

本文通过药学部1984～1994年间共10年的医院制剂生产利润、全院处方金额、平均每张处方金额及处方数分别经最小二乘法处理,认为前三者与年度呈线性关系,并对1994年的资料进行了预测,认为药学部94年制剂及调剂收入情况在93年基础上有所增长,平均每张处方金额与93年持平。     

基于最小二乘法的艾滋病疗法疗效预测

根据美国最大的艾滋病医疗试验机构ACTG公布的一组数据来对AIDS疗法的疗效进行预测。以MATLAB为平台,利用最小二乘法分别得到病人的CD4和HIV浓度总的平均变化量与时间t的拟合曲线。由拟合曲线可得到在40周时CD4浓度总的平均增加量达到最大,在14周半时而HIV浓度总的平均减少量达到了最大。综合这两个时间可得到当这300多名病人同时服用齐多夫定,拉美夫定和茚地那韦3种药物进行AIDS治疗时,应该在27周左右时终止治疗。     

偏最小二乘法用于近红外漫反射光谱快速测定异福片

目的:采用近红外光谱技术,建立一种快速测定异福片中利福平和异烟肼含量的方法。方法:将光谱分别进行卷积平滑、一阶导数、二阶导数预处理,应用偏最小二乘法(PLS)建立定量分析模型,以校正集的交互验证均方根误差(RMSECV)及相关系数(r)为优化参数,选择最佳预处理方法和波长范围。通过留一交互验证法,以预测残差平方和(PRESS)为优化参数,选择最适主因子数。结果:校正集样品利福平和异烟肼的 r 分别为0.9940和0.9910,RMSECV 分别为0.00764和0.00445。对预测集样品利福平和异烟肼的预测均方根误差(RMSEP)分别为0.00397和0.00383;平均加样回收率分别为100.88%和100.87%;重复性试验 RSD 分别为0.672%和1.08%。结论:结果表明该方法预测精度高,且具有方便快捷、非破坏、无污染、可在线检测、重复性好等优点,可作为异福片原位质量检测和在线质量监控的方法予以推广。     

最小二乘法在卫生理化检验中的应用简介

论述最小二乘法的原理及其在组合测量、线性回归等方面的应用。     

基于偏最小二乘模型的奥氮平血药达峰浓度与个体内在因素的相关性研究

目的:探讨奥氮平血药达峰浓度(Cmax)与个体内在因素的相关性.方法:采用LC-MS/MS测定20名健康志愿者奥氮平的Cmax,采用偏最小二乘回归方法(PLS)分析Cmax与个体内在因素包括年龄、体重指数(BMI)、白细胞(WBC)、红细胞(RBC)、血小板(PLT)、血红蛋白(HB)、肌酐(Cr)、尿素氮(BUN)、总蛋白(TP)、白蛋白(ALB)、碱性磷酸酶(AKP)、谷丙转氨酶(ALT)等的相关性.结果:通过PLS分析结果表明:Cmax与BMI、WBC、BUN、TP、ALT呈正相关;与RBC、PLT、HB、Cr、ALB、AKP呈负相关,其中与BMI、WBC、RBC、BUN、TP相关性最强.预测建模分析显示,模型拟合效果较好,可以作为预测奥氮平体内Cmax的参考指标.结论:个体内在因素影响奥氮平的Cmax,可以为临床个体化用药提供参考.     

基于偏最小二乘法关联分析栀子不同炮制品化学成分与肝肾毒性

目的 基于偏最小二乘回归法（PLS）研究栀子不同炮制品化学成分与肝肾毒性的关联度。方法 采用超高效液相色谱法测定生栀子、炒栀子、姜栀子、焦栀子、栀子炭不同炮制品中栀子苷、京尼平龙胆双糖苷、栀子苷酸、去乙酰车叶草酸甲酯、西红花苷I和西红花苷Ⅱ 6种化学成分的含量,栀子不同炮制品水提液（7.5 g/kg）连续3 d大鼠ig给药,测定各组大鼠血清丙氨酸氨基转移酶（ALT）、天冬氨酸氨基转移酶（AST）、总胆红素（TBIL）、血清肌酐（CREA）、尿素氮（BUN）的含量,使用Origin 9.1软件中的PLS算法分析栀子不同炮制品化学成分与肝肾毒性的关联度。结果 栀子炮制后栀子苷、西红花苷I和西红花苷Ⅱ的含量不同程度的降低,京尼平龙胆双糖苷、去乙酰车叶草酸甲酯和栀子苷酸含量在部分炮制品中升高;栀子及炮制品对正常大鼠可造成不同程度的肝、肾损伤（生栀子>炒栀子>姜栀子>焦栀子>栀子炭）;PLS分析表明6种主要化学成分中栀子苷与西红花苷I的变量重要性值均超过0.8。结论 栀子苷与西红花苷I可能是栀子肝肾毒性的物质基础,通过炮制降低栀子肝肾毒性的作用可能与炮制后栀子苷与西红花苷I含量降低有关。     

Population Pharmacodynamic Parameter Estimation from Sparse Sampling: Effect of Sigmoidicity on Parameter Estimates

The objective of this stimulation study was to evaluate effect of simoidicity of the concentration–effect (C–E) relationship on the efficiency of population parameter estimation from sparse sampling and is a continuation of previous work that addressed the effect of sample size and number of samples on parameters estimation from sparse sampling for drugs with C–E relationship characterized by high sigmoidicity (γ > 5). The findings are based on observed C–E relationships for two drugs, octreotide and remifentanil, characterized by simple E_max and sigmoid E_max models (γ = ~2.5), respectively. For each model, C–E profiles (100 replicates of 100 subjects each) were simulated for several sampling designs, with four or five samples/individual randomly obtained from within sampling windows based on EC₅₀-normalized plasma drug concentrations, PD parameters based on observed population mean values, and inter-individual and residual variability of 30% and 25%, respectively. The C–E profiles were fitted using non-linear mixed effect modeling with the first-order conditional estimation method; variability parameters were described by an exponential error model. The results showed that, for the sigmoid E_max model, designs with four or five samples reliably estimated the PD parameters (EC₅₀, E_max, E₀, and γ), whereas the five-sample design, with two samples in the 2–3 E_max region, provided in addition more reliable estimates of inter-individual variability; increasing the information content of the EC₅₀ region was not critical as long as this region was covered by a single sample in the 0.5–1.5 EC₅₀ window. For the simple E_max model, because of the shallower profile, enriching the EC₅₀ region was more important. The impact of enrichment of appropriate regions for the two models can be explained based on the shape (sigmoidicity) of the concentration–effect relationships, with shallower C–E profiles requiring data enrichment in the EC₅₀ region and steeper curves less so; in both cases, the E_max region needs to be adequately delineated, however. The results provide a general framework for population parameter estimation from sparse sampling in clinical trials when the underlying C–E profiles have different degrees of sigmoidicity.Key words: clinical trials, PK-PD simulations, population PK-PD modeling, sparse sampling     

基于神经网络的项目参数估计方法

针对题库建设中项目参数估计的实际问题,提出了一种全新的基于神经网络的参数估计方法;并以二值记分的3PLM为项目反应理论模型,以广义回归神经网络为网络模型,根据Monte Cado实验法进行了模拟实验研究,最后将神经网络方法与传统的数理统计估计方法进行了比较.结果表明,在小样本测验情况下,神经网络方法具有一定的优势,尤其是当去掉对项目参数的先验概率分布的限制时,神经网络方法的优势更加明显,说明本文提出的方法具有一定的价值. ,     

Experimental Design and Efficient Parameter Estimation in Preclinical Pharmacokinetic Studies

Monte Carlo simulation technique used to evaluate the effect of the arrangement of concentrations on the efficiency of estimation of population pharmacokinetic parameters in the preclinical setting is described. Although the simulations were restricted to the one compartment model with intravenous bolus input, they provide the basis of discussing some structural aspects involved in designing a destructive (quantic) preclinical population pharmacokinetic study with a fixed sample size as is usually the case in such studies. The efficiency of parameter estimation obtained with sampling strategies based on the three and four time point designs were evaluated in terms of the percent prediction error, design number, individual and joint confidence intervals coverage for parameter estimates approaches, and correlation analysis. The data sets contained random terms for both inter- and residual intra-animal variability. The results showed that the typical population parameter estimates for clearance and volume were efficiently (accurately and precisely) estimated for both designs, while interanimal variability (the only random effect parameter that could be estimated) was inefficiently (inaccurately and imprecisely) estimated with most sampling schedules of the two designs. The exact location of the third and fourth time point for the three and four time point designs, respectively, was not critical to the efficiency of overall estimation of all population parameters of the model. However, some individual population pharmacokinetic parameters were sensitive to the location of these times.     

Estimation of the Probability of Passing the USP Dissolution Test

To ensure that a drug product will meet standards for identity, strength and stability as specified in the United States Pharmacopedia and National Formulary (USP/NF), it needs to pass a number of tests such as the content uniformity test and dissolution test at various stages of the manufacturing process. The sponsors usually have in-house specification limits based on some lower bounds of the probabilities of passing USP/NF tests to make sure that there is a high probability of passing the tests. Several probability lower bounds for dissolution test have been provided in the literature. In this paper, a method of calculating the probabilities of passing the dissolution tests is proposed. For the population mean and variance in some specified range, the probability derived from the methodology is very close to the exact probability. Therefore, the proposed method can provide an easy and accurate way to calculate the probability.     

移动平均线数学模型的参数估计和预测

建立了二次移动平均线数学模型 ,对模型中参数进行了估计和检验 ,并对模型进行了预测     

一类容易出错的参数估计问题

在用置信区间解决某些实际问题时,若感兴趣的只是置信区间的上限或下限,此时只需求出单侧置信区间即可。     

Similarity or Discrepancy in Pharmacokinetic Parameter Estimation Between Bolus and Infusion Studies

A recent study by Heatherington and Rowland showing discrepancies in steady-state volume of distribution (V_ss) estimation of two barbiturates between bolus and infusion studies in rat hindlimb preparations was reviewed. Their rationale is that increasing the duration of administration may increase the accessibility for tissue distribution and thus increase V_ss for compounds showing slow tissue uptake. Such a dosing-duration-dependent distribution concept is, however, inconsistent with the principle in linear kinetics that the fate or disposition function of any drug molecules is independent of time of administration and presence of other molecules. When their well-designed bolus studies were reanalyzed by including extrapolated outflow data from the last sampling time to infinity, the V_ss values for the two barbiturates were found to be very similar to those obtained by the infusion method. Our analysis seems to validate a theoretical concept that parameter estimation is independent of the duration of administration in linear kinetics. A potential complication of using the bolus method to study V_ss is presented.     

Estimation of Dissolution Rate From In Vivo Studies of Synthetic Vitreous Fibers

Although the dissolution rate of a fiber was originally defined by a measurement of dissolution in simulated lung fluid in vitro, it is feasible to determine it from animal studies as well. The dissolution rate constant for a fiber may be extracted from the decrease in long fiber diameter observed in certain intratracheal instillation experiments or from the observed long fiber retention in short-term biopersistence studies. These in vivo dissolution rates agree well with those measured in vitro for the same fibers. For those special types of fibers, the high-alumina rock wool fibers that could not be measured in vitro, the method provides a way of obtaining a chemical dissolution rate constant from an animal study. The inverse of the in vivo dissolution rate, the fiber dissolution time, correlates well with the weighted half life of long fibers in a biopersistence study, and the in vivo dissolution rate may be estimated accurately from this weighted half-life.