hce package. The standardized output object is now
of class adhce. This means also that there is a more standardized output
that make it easier to work with. As a result, the maraca package has now
a higher hce version dependency (0.8.5). Also, all class dependent functions
in the package have been updated to only work on the adhce class object
(for example plot.adhce()).remove_outliers in plot.maraca() and plot.adhce().
In some cases, there might be outliers that skew the displayed range for the
continuous endpoint. There is now an option to display the continuous
endpoint without the outliers by setting the parameter
remove_outliers = TRUE in the plot.maraca() or plot.adhce() function.
We define outliers here according to the common boxplot calculation
definition: any data outside the range 25th percentile - 1.5 * IQR
(inter-quartile range) and 75th percentile + 1.5 * IQR.
Note that this required some refactoring in how the plot is constructed.
Especially the violin plot is now pre-calculated and then plotted using the
ggplot2 function geom_polygon() (rather than the geom_violin()
function).density_plot_type = "box" is selected, the boxplot will now
contain vertical segments to indicate where the whiskers end.animate_maraca()
function. This is an animated version of the standard maraca plot to
allow to show how the plot is being built up step-by-step. Note that the
gganimate package needs to be installed to create the animation.
Additionally, to save the animation as a gif, the package gifski needs
to be installed.
This is an experimental feature, so despite doing some testing during
development there might be some problems or unexpected behavior during
usage. Please take a minute to report any wrong behavior to allow us to
improve the functionality.Slight change in automatic checks after an update of the hce package
(dependency).
mosaic_plot - a new plot to compares outcome between an
active treatment group and a control group, highlighting areas of "Wins",
"Losses" and "Ties" based on endpoint hierarchy. Details are given in
the new vignette "Maraca Plots - Introduction to the Mosaic plot".cumulative_plot() function - dustin() and
dustin_plot().Updated author information.
component_plot(), there has been a new plot added called
cumulative_plot(). As opposed to the previous plot showing the individual
components of the win odds computation, this plot is displaying
the endpoints cumulated instead (adding one component of hierarchical endpoint
at a time). Details can be found in the vignette "Maraca Plots - Plotting win odds".tte_outcomes has been changed to step_outcomes and the parameter
continuous_outcome to last_outcome.ggplot2 is now automatically attached when loading maraca.maraca has a new dependency - the patchwork package.trans parameter in the plotting functions was not working as
intended. It now enables x-axis transformation for the continuous
endpoint part of the plot.theme argument in the plotting functions allows users to easily change the
styling of the plot. Details are given in the new vignette
"Maraca Plots - Themes and Styling".component_plot()
function works for both maraca and hce. Details can be found in the new
vignette "Maraca Plots - Plotting win odds".validate_maraca() that was added in version 0.5.maraca now has increased the version dependency for the package hce
to >= 0.5.hce package is now automatically attached when loading maraca.print() function for maraca objects that summarizes key information.validate_maraca() function that extracts key information from a maraca
plot object. This can be used to validate the plot against independently coded
versions (for example using a different programming language).maraca() function now requires an input for the parameter
fixed_followup_days. Note that there can be no observed events in the
data after the follow-up time specified.maraca does no longer depend on the gridExtra package.plot_tte_components() function for plotting the individual time-to-event
outcomes was removed from the package since it did not prove to be overly
useful.plot_tte_composite() was removed for now since the package cannot correctly
calculate the composite version of looking at multiple time-to-event endpoints when
patients have multiple events.