Using Pharsight® Trial Simulator™

Guided Tour

Pharsight Trial Simulator uses Monte Carlo simulations, based on a stochastic model of drug action, to approximate the distribution of probable outcomes for a clinical trial. Batch simulations can include multiple variations on the model and trial design, as well as statistical analysis of each variation. The process can be driven from Visual Basic or another similar scripting language. Thus, the Trial Simulator makes it easy to compare trial designs and to evaluate the influence of assumptions about how subjects will behave in these trials. For example, to test different designs, you might compare competing observation schedules or change the enrollment size in the study protocol. To test different model assumptions, you might vary key pharmacodynamic parameters across their confidence range, or compare different models of patient compliance.

Each simulation is repeated many times to estimate the distribution of possible outcomes for each variation on the model and trial. Many simulations, or replicates, of many study variations can be simulated in one batch in Trial Simulator, so variations on a trial are easily compared.

Workflow in the Trial Simulator

A simulation project proceeds in two stages. The first stage involves getting the software to represent the problem to be examined. This is very much like a programming task. You "code" the model and study in the various graphical editors in Trial Simulator, then run the simulation to confirm that it reflects your intentions.

Once you are confident that the Trial Simulator represents your drug and patients, the focus changes to using the software to investigate the problem. This investigation involves exploring two types of information: alternative assumptions about how subjects will respond and different trial designs. More...

The Trial Simulator Application Window

The Trial Simulator window organizes your simulation project into modules, accessible through a left-side tree control or main study diagram. Simulation progress appears in a separate simulation window. More...

Trial Simulator Studies

Each Trial Simulator project is called a study. A study includes one main model, protocol and analysis plan, as well as variations on the main model and protocol, and simulation results. More...

The study elements are organized into seven major sections:

    Covariate Distribution Model: demographic information for one or more sub-populations. More...

    Drug Model: stochastic model of drug actions in individual subjects. More...

    Nominal Design: the planned protocol. More...

    Protocol Deviations: (optional) model of non-adherence to the nominal protocol. More...

    Data Analysis Plan: (optional) statistical analysis of each study replicate. More...

    Simulation Scenarios: comparing variations on the drug model and protocol. More...

    Results: simulation database analysis and export to compare scenarios More...