A major goal of research in Clinical Pharmacology is to characterize and understand the quantitative relationships between drug exposure and drug effects. While research questions largely justify this goal, there is also a more practical motivation: to optimally individualize the use of effective but toxic drugs requires considerable knowledge of the relationship between drug exposure, response(s) and patient characteristics, the so-called response surface. Model-based analysis. of PK/PD data arising from many sources, including non-classical studies, is increasingly seen as a means to gain knowledge about the response surface, and thence to better dosage and better clinical outcomes. The data analysis methods we have studied and plan to continue to study provide the inferential tools that are necessary for analyses of such data. During the requested continuation period we specifically propose to address the following areas: 1. Analysis of Population PK/PD Data. We plan to further develop and study an improved (less biased) estimation method than our original method, the First Order (FO) method, the so-called First Order Conditional Estimate (FOCE) method. We will pay particular attention to applying it to (i) categorical, mixed categorical/continuous, and survival-time type data (as arise in PD, as opposed to PK applications); (ii) data arising from asymmetric distributions; and (iii) multivariate/multiple response data. 2. Designs for Population PK/PD Studies. We propose to study PK/PD design strategies under 4 analysis approaches: either combine the PK/PD data or analyze it separately; use the FO or FOCE method. We will examine the design implications of PK and PD model misspecification and errors in independent variables (dose/sample timing; dose magnitude). 3. Implementation. Several additional functionalities will be added to NONMEM: (i) the new methods developed in aim #1, and (ii) the ability to fit models with 3 levels of random effects (rather than the current 2.) This will enable fitting models with """"""""observation"""""""" error, intra-individual (day-to-day) kinetic variability and inter-individual variability (important, e.g., in bioequivalence studies). Along with the above coding efforts, we plan to develop updated and additional documentation aids, notably to improve our 'on-line' help document that should, eventually, provide a reference to all NMTRAN, PREDPP, and NONMEM key words, plus control-stream examples of the use of each.
Showing the most recent 10 out of 55 publications
