Skip to contents

PK Parameters

PK_Parameter.Rmd

Introduction

This vignette provides a comprehensive overview of how to calculate and visualize pharmacokinetic (PK) parameters across various scenarios using the ospsuite.reportingframework package. This guide will walk you through the process of utilizing the PKParameter.xlsx configuration table, importing observed data, and generating insightful visualizations to compare simulated and observed outcomes.

Calculate PK Parameters for Scenarios

The PK-parameter calculation is an extension of the esqlabR package. A new configuration table, PKParameter.xlsx, can be accessed via projectConfiguration$addOns$pKParameterFile.

This table contains a template sheet listing all available PK parameters in the OSPSuite, along with examples of two user-defined PK parameters: F_max and F_end. User-defined PK parameters can be specified in the Userdef PK Parameter sheet.

Internally, the calculations rely heavily on the functions ospsuite::updatePKParameter() and ospsuite::addUserDefinedPKParameter(). If any aspect is unclear, please refer to the respective function documentation for detailed information.

To configure the PK-parameter selection for a scenario:

  1. Copy the “Template” sheet.
  2. Delete any unnecessary rows.
  3. Add any missing user-defined parameters.

The projectConfiguration$scenariosFile table, as an extension of the esqlabR package, includes an additional sheet called ‘PKParameter’. Here, you can link each scenario to the corresponding sheet name from the PKParameter.xlsx.

When calling runOrLoadScenarios or runAndSaveScenarios, the selected PK parameters are automatically calculated and saved in the PKAnalysisResults folder within the output directory. If simulation results already exist and you wish to recalculate the PK parameters without initiating a new simulation, use calculatePKParameterForScenarios.

To load the PK parameters use:

pkParameterDT <- loadPKParameter(
  projectConfiguration = projectConfiguration,
  scenarioListOrResult = scenarioResults
)

Import Observed Data

If available, you can to import the observed PK data. This can be accomplished using a data.table containing the relevant observed parameters, which can then be compared against simulated data. The data import process is described in detail in the vignette Data_import_by_dictionary.

Available Plot Types

Overview of Plot Types

Plot Type Objective Simulated Data Observed Data Numeric Values Data Types
Box-whisker plots Display and compare distributions of different scenarios. Displayed are percentiles as box-whiskers. Tables containing the numeric values are exported additionally to each plot as CSV. Absolute values and ratios of crossover studies.
Forest plots of aggregated distributions Compare aggregated distributions of different scenarios. Compare observed data with simulated values. Displayed is point and range (e.g., geometric mean and geometric standard deviation). Aggregated data can be added as point range. Tables containing the numeric values are optionally displayed. Absolute values and ratios of crossover studies.
Forest plots of point estimators Compare point estimators (e.g., DDI Ratios) of different scenarios. Compare observed data with simulated values. Estimator is displayed as a point. Aggregated data can be added as point range. Tables containing the numeric values are optionally displayed. Absolute values, ratios of crossover studies, and ratios of scenarios if different studies.
Histograms Display and compare distributions of different scenarios. Displayed is the distribution as a histogram. Absolute values and ratios of crossover studies.
Range plots Display dependence of PK parameters vs. population parameters (e.g., age). Displayed is aggregated distribution per bin of population parameter. Exported are the aggregated values per bin. Absolute values and ratios of crossover studies.

Please check the corresponding function help for detailed information on each plot type and the vignette Plot and Report Generation.

Box-whisker plots

Box-whisker plots provide a visual representation of the distribution of PK parameters across different scenarios. Default percentiles are set to 0.05, 0.25, 0.5, 0.75, and 0.95. Data outliers can be added as points. Additionally, tables containing the number of individuals, percentiles, mean, standard deviation (SD), geometric mean, geometric SD, and geometric coefficient of variation (CV) of the simulated data are exported alongside each plot.

To generate these plots, use the plot function plotPKBoxwhisker as input for the runPlot function.

Forest plots of aggregated distributions

Forest plots allow for the comparison of aggregated distributions of different scenarios. The estimator for mean and range can be selected, with available options including arithmetic mean and standard deviation; geometric mean and standard deviation; and custom functions. This functionality is particularly useful for visualizing how simulated data aligns with observed data, enhancing the interpretability of PK parameter estimates.

To generate these plots, use the plot functions plotPKForestAggregatedAbsoluteValues or plotPKForestAggregatedRatios as input for the runPlot function.

Forest plots of estimators

Forest plots of estimators provide a means to compare point estimators (e.g., DDI Ratios) across different scenarios. These plots can include observed data, allowing for a direct comparison between simulated and actual outcomes. The estimators are displayed as points, with confidence intervals represented as ranges. The confidence interval reflects the uncertainty based on the number of simulated individuals, not the model uncertainty.

To generate these plots, use the plot functions plotPKForestPointEstimateOfAbsoluteValues or plotPKForestPointEstimateOfRatios as input for the runPlot function.

Histograms of distributions

Histograms display the distribution of PK parameters for different scenarios. By visualizing the frequency of parameter values, histograms provide insights into the underlying distribution shapes, such as normality or skewness, which are critical for statistical analysis and interpretation.

To generate these plots, use the plot function plotHistograms as input for the runPlot function.

Range plots

Range plots illustrate the dependence of PK parameters on population parameters, such as age. Aggregated distributions are displayed per bin of population parameters, allowing for the identification of trends and patterns. Exported are the aggregated values per bin, which can be useful for further statistical analysis and reporting. Data types include both absolute values and ratios of crossover studies, providing a comprehensive view of the relationships between parameters.

To generate these plots, use the plot function plotDistributionVsDemographics as input for the runPlot function.