Skip to contents

Path data class (R6 class)

PathData.Rd

It is not recommended for users to try to create data using the constructor of this class. Rather use df_to_pathdata.

Public fields

subject_df

Data frame with one row per subject. Must have one row for each subject, and subject_id and all covariates as columns.

path_df

Data frame of actual paths. Each row corresponds to one time point. Must have path_id, state, time, and trans_idx as columns. These are

  • path_id: path identifier

  • state: integer index of the state at the time point, and until the next time point

  • time: time point

  • trans_idx: Integer indicating which transition this time point corresponds to. If the time point is not a state transition, this should be 0.

link_df

Links the path and subject data frames. Must have one row for each path, and path_id, draw_idx, rep_idx, and subject_id as columns.

covs

Covariate column names.

transmat

A TransitionMatrix describing the system to which the paths belong.

Methods

Public methods

Method new()

Initialize

Usage

PathData$new(subject_df, path_df, link_df, transmat, covs = NULL)

Arguments

subject_df

Data frame with one row per subject. Must have one row for each subject, and subject_id and all covariates as columns.

path_df

Data frame of actual paths. Each row corresponds to one time point. Must have path_id, state, time, and trans_idx as columns. These are

  • path_id: path identifier

  • state: integer index of the state at the time point, and until the next time point

  • time: time point

  • trans_idx: Integer indicating which transition this time point corresponds to. If the time point is not a state transition, this should be 0.

link_df

Links the path and subject data frames. Must have one row for each path, and path_id, draw_idx, rep_idx, and subject_id as columns.

transmat

A TransitionMatrix describing the system to which the paths belong.

covs

Covariate column names.

Method unique_subjects()

Get unique subject ids

Usage

PathData$unique_subjects()

Method covariate_names()

Get names of covariates

Usage

PathData$covariate_names()

Returns

a character vector

Method subset_covariates()

Create a new PathData with a subset of covariates

Usage

PathData$subset_covariates(covs, renamed_old = NULL, renamed_new = NULL)

Arguments

covs

Names of all new covariates

renamed_old

Name of an old covariate to rename

renamed_new

Name of new covariate where old one is copied

Method state_at()

For each path, get the state it is in at time t

Usage

PathData$state_at(t)

Arguments

t

time

Returns

a data frame with one row for each path

Method transition_times()

Get all transition times

Usage

PathData$transition_times(trans_inds = NULL)

Arguments

trans_inds

Indices of transition whose occurrence times are requested. Default is NULL, in which case all transitions are considered.

Returns

a numeric vector

Method n_paths()

Get number of paths

Usage

PathData$n_paths()

Returns

an integer

Method state_names()

Get names of all states

Usage

PathData$state_names()

Returns

character vector

Method get_event_state_names()

Get names of all states to which transitioning can be a transition event

Usage

PathData$get_event_state_names()

Returns

character vector

Method terminal_states()

Get names of terminal states

Usage

PathData$terminal_states()

Returns

character vector

Method print()

Print info

Usage

PathData$print()

Returns

nothing

Method get_path_df()

Get path data frame

Usage

PathData$get_path_df(truncate = FALSE)

Arguments

truncate

Remove rows after terminal events?

Method as_data_frame()

Convert to one long data frame

Usage

PathData$as_data_frame(covariates = NULL, truncate = FALSE)

Arguments

covariates

Which covariates to include?

truncate

Remove rows after terminal events?

Method full_link()

Full link data frame

Usage

PathData$full_link(covariates = NULL)

Arguments

covariates

Which covariates to include

Returns

A data.frame with same number of rows as link_df, including also the covariate columns and subject_id

Method as_transitions()

Data frame in transitions format

Usage

PathData$as_transitions(covariates = NULL, truncate = FALSE)

Arguments

covariates

Which covariates to include?

truncate

Remove rows after terminal events first?

Returns

A data.frame

Method as_transitions_alt()

Data frame in alternative transitions format

Usage

PathData$as_transitions_alt(covariates = NULL, truncate = FALSE)

Arguments

covariates

Which covariates to include?

truncate

Remove rows after terminal events first?

Returns

A data.frame

Method as_msdata()

Convert to format used by the 'mstate' package

Usage

PathData$as_msdata(covariates = NULL)

Arguments

covariates

Which covariates to include?

Returns

An msdata object

Method plot_paths()

Step plot of the paths

Usage

PathData$plot_paths(n_paths = NULL, alpha = 0.5, truncate = FALSE)

Arguments

n_paths

Number of paths to subsample for plotting.

alpha

opacity

truncate

truncate after terminal events?

Method prop_matrix()

Transition proportion matrix

Usage

PathData$prop_matrix(include_censor = TRUE)

Arguments

include_censor

Include censoring (no event) proportion in the matrix

Returns

a table

Method plot_graph()

Visualize the transition proportion matrix as a graph

Usage

PathData$plot_graph(digits = 3, ...)

Arguments

digits

Max number of digits to show in numbers

...

Arguments passed to qgraph

Returns

qgraph plot

Method fit_coxph()

Fit Cox proportional hazards model

Usage

PathData$fit_coxph(covariates = NULL, ...)

Arguments

covariates

Covariates to include.

...

Arguments passed to survival::coxph.

Method fit_mstate()

Fit frequentist 'mstate' model

Usage

PathData$fit_mstate(covariates = NULL, ...)

Arguments

covariates

Covariates to include.

...

Arguments passed to survival::coxph.

Method filter()

Filter based on subject id, creates new object

Usage

PathData$filter(subject_ids_keep)

Arguments

subject_ids_keep

Subject ids to keep

Method clone()

The objects of this class are cloneable with this method.

Usage

PathData$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.