FLEX-FIT: A Flexible Bayesian Forecasting Program for Dosage Optimization in Clinical Pharmacokinetics

Abstract

A microcomputer program for prediction of individualized optimal drug dosage based on Bayesian Statistical Theory and Maximum Likelihood Estimation is presented. Model (one or two compartmental model) and parameter distribution (normal or log normal) can be selected to describe the plasma drug concentrations-time course. Simplex method is adapted for contraction of objective function in non-linear least square fitting to the final minimum. It is designed so that user can integrate universal dosing regimens with different doses, routes, and regular or irregular intervals. Complex dosage regimens and non steady-state conditions can be handled. This program can also generate graphic simulations of plasma and peripheral compartment drug levels of one or two compartmental model. It has been designed to run interactively to assist learning the pharmacokinetic concept and to be handled by user with little knowledge on mathematics or computer.