Skip to contents

Bayesian multistate modeling of right-censored time-to-event data.

Transition matrix

For any analysis with this package, you will need a TransitionMatrix which specifies the states and possible transitions between them. This can be created directly with TransitionMatrix$new() or using built-in functions for common multistate systems such as transmat_illnessdeath.

Data format

See the function df_to_pathdata for how to create a PathData object from your data frame and a given TransitionMatrix.

Creating a model

You can create a model using create_msm. A model is represented by a MultistateModel object, which has a MultistateSystem, which has a TransitionMatrix.

Fitting a model

You can fit a model with 'Stan' using fit_stan.

Prediction with a fitted model

Vignettes

See the vignettes for more information.

See also

Author

Juho Timonen and Jacqueline Buros