R Jelliffe, A Schumitzky, D Bayard, R Leary, M Van Guilder, S Goutelle, A Bustad, A Botnen, J Bartroff, W Yamada, and M Neely
Laboratory of Applied Pharmacokinetics, USC Keck School of Medicine, Los Angeles, CA, USA
Obtectives: A current limitation of IMM Bayesian sequential analysis [1] is that the changing parameter distributions which take place during the period of data analysis may lie in a region of the nonparametric population model which has few support points.
Methods: To improve this situation. our group has added extra support points from the augmented population model currently used for our hybrid Bayesian analysis of fixed parameter distributions during the data analysis The hybrid procedure starts with a maximum aposteriori probability Bayesian analysis. Extra support points were added near the MAP Bayesian estimate to form a 4 x 4 grid of support points. This has now been upgraded to 10 x 10 points, with an adjustable percent change between them, ranging from 5 to 20 percent.
Results: This augmented population model now provides a much larger set of points for the changing parameter values of the unstable patient to attach to, and provides a richer set of support points from which to compute the next adjustment of the dosage regimen to achieve the desired target goal most precisely, with minimum expected weighted squared error.
Conclusions: Past versions of this software have tracked drug behavior in patients better that other Bayesian methods [2]. This new improved population model prior enhances this capability.
References:
[1] Bayard D, and Jelliffe R: A Bayesian Approach to Tracking Patients having Changing Pharmacokinetic Parameters. J. Pharmacokin. Pharmacodyn. 31 (1): 75-107, 2004.
[2] Macdonald I, Staatz C, Jelliffe R, and Thomson A: Evaluation and Comparison of Simple Multiple Model, Richer Data Multiple Model, and Sequential Interacting Multiple Model (IMM) Bayesian Analyses of Gentamicin and Vancomycin Data Collected From Patients Undergoing Cardiothoracic Surgery. Ther. Drug Monit. 30:67-74, 2008
Reference: PAGE 21 () Abstr 2600 [www.page-meeting.org/?abstract=2600]
Poster: Estimation methods