Jon Wakefield
Imperial College, London, UK
At last year’s American Statistical Association’s meeting in Boston an organised session compared various approaches to population pharmacokinetic analyses, basing the comparison on data for the antiarrhythmic drug quinidine. In this talk I shall describe and compare the methodology behind the various approaches and summarise the findings of the analyses. In particular I shall discuss the NONMEM approach that was carried out in the original analysis (Verme et al, 1992), the semiparametric approach of Davidian and Gallant (1993), the nonparametric approach of Mallet (1993) and my own analysis using a parametric Gibbs sampling approach, Wakefield (1993).
The dataset consisted of dosing histories, concentration/time data and nine covariate measurements for 136 hospitalized males. There were a total of 361 serum quinidine measurements in total. The main aim of the analysis was the determination of those covariates that are important for the prediction of clearance.
References:
[1]. Davidian. M. and Gallant, A.R. (1993). Smooth nonparametric maximum likelihood estimation for population pharmacokinetics, with application to quinidine. J. Pharmacok. Biopharm. 20, 529-556.
[2]. Mallet. A. (1993). A nonparametric approach for population pharmacokinetics. Submitted for publication.
[3]. Verme. C.N., Ludden, T.M., Clementi, W.A. and Harris, S.C. (1992). Pharmacokinetics of quinidine in male patients. Clin. Pharmacokin. 22, 468-480.
[4]. Wakefield. J.C. (1993). The Bayesian analysis of population pharmacokinetic models. Submitted to Journal Of The American Statistical Association.
Reference: PAGE 2 (1993) Abstr 903 [www.page-meeting.org/?abstract=903]
Poster: oral presentation