| With the continuous development of the theory and research of individualized drug delivery,more and more vancomycin dose-assisted decision-making tools have been developed and applied.However,the pharmacokinetic model,blood concentration estimation method and functional range of different tools are different,so the clinical applicability and prediction accuracy of each tool in Chinese patients are still unclear.Based on this,we designed a Gibbs sampling scheme with adjustable prior intensity based on PPK model and applied it to vancomycin individualized administration scheme to obtain more accurate prediction of individual PK parameters and to predict drug dosage.First,the objective of this study was to design an efficient Gibbs sampling protocol based on Bayesian and population pharmacokinetic models for use in adults(including the elderly)treated with vancomycin,combined with simulated data on the biological characteristics of the population,and ultimately to determine the optimal dose for individuals.And can effectively give a reasonable dose range and confidence level.Secondly,serum creatinine(SCr),the index that has the greatest influence on patients’ blood drug concentration,was used to conduct numerical simulation,and the PPK model for target population was constructed by using the known nonlinear mixed effects(NONMEN)model.A method of Gibbs sampling which can adjust the prior intensity and Gibbs sampling for the error terms was designed.The optimal PK parameters and dosage of the individual were investigated.This process is consistent with the whole process of individualized drug use.Finally,the model effect of the simulation results is evaluated.In the evaluation of 100 patients,the mean prediction error(MPE)and root mean square error(RMSE)of observed and measured blood concentration were 1.0828 mg·L-1 and 1.1639 mg·L-1,respectively.The residual error of the model obeyed the standard normal distribution,showing a reasonable scheme and good effect.In addition,in the six kinds of convergence diagnosis for the sampled samples of a single individual,there is convergence in the sample chain.In addition,the characteristics of the patient with the best predictive effect were also analyzed.The results show that this sampling scheme has important enlightening value for individualized drug delivery with known NONMEN’s PPK model. |
