Pharmacometrics: modelling and simulation tools to improve decision making in clinical drug development.

摘要: There is broad recognition within the pharmaceutical industry that drug development process, especially clinical part of it, needs considerable improvement to cope with rapid changes in research and health care environments. Modelling simulation are mathematically founded techniques have been used extensively for a long time other areas than (e.g. automobile, aerospace) design develop products more efficiently. Both modelling rely on use (mathematical statistical) models which essentially simplified descriptions complex systems under investigation. It has proposed integrate pharmacokinetic (PK) pharmacodynamic (PD) principles into make it rational efficient. evidence from survey 18 projects PK/PD guided approach can contribute streamline process. This relies describing relationships among dose, concentration (and generally exposure), responses such as surrogate markers, efficacy measures, adverse events. Well documented empirical physiologically based becoming available more, there ongoing efforts disease progression patient behavior compliance) well. Other types increasingly important population which, addition characterization PK PD, involve between covariates (i.e. characteristics age, body weight) parameters. Population allow assess quantify potential sources variability exposure response target population, even sparse sampling conditions. As will be shown an anticancer agent, implications significant covariate effects evaluated by computer simulations using model. Stochastic widely tool evaluation statistical methodology including example performance measures bioequivalence assessment. Recently, was suggested expand support predicting outcomes planned trials. The methodological basis this provided (population) together random techniques. Models behavioral features like compliance, drop-out rates, event dependent dose reductions, etc. added order mimic real situation. helps evaluate consequences safety assessment drug, enabling identification statistically valid practically realisable study designs. For both guidance 'best practices' currently worked out panel experts comprising representatives academia, regulatory bodies industry, thereby providing necessary condition model-based analysis further streamlining processes.