Package: TimeDepFrail 0.0.1
Alessandra Ragni
TimeDepFrail: Time Dependent Shared Frailty Cox Model
Fits time-dependent shared frailty Cox model (specifically the adapted Paik et al.'s Model) based on the paper "Centre-Effect on Survival After Bone Marrow Transplantation: Application of Time-Dependent Frailty Models", by C.M. Wintrebert, H. Putter, A.H. Zwinderman and J.C. van Houwelingen (2004) <doi:10.1002/bimj.200310051>.
Authors:
TimeDepFrail_0.0.1.tar.gz
TimeDepFrail_0.0.1.tar.gz(r-4.5-noble)TimeDepFrail_0.0.1.tar.gz(r-4.4-noble)
TimeDepFrail_0.0.1.tgz(r-4.4-emscripten)TimeDepFrail_0.0.1.tgz(r-4.3-emscripten)
TimeDepFrail.pdf |TimeDepFrail.html✨
TimeDepFrail/json (API)
| # Install 'TimeDepFrail' in R: |
| install.packages('TimeDepFrail', repos = 'https://cloud.r-project.org') |
Bug tracker:https://github.com/alessandragni/timedepfrail/issues
Pkgdown site:https://alessandragni.github.io
- data_dropout - Data Dropout Dataset
Last updated 3 months agofrom:d4dff12b3d. Checks:3 OK. Indexed: no.
| Target | Result | Latest binary |
|---|---|---|
| Doc / Vignettes | OK | Mar 19 2025 |
| R-4.5-linux | OK | Mar 19 2025 |
| R-4.4-linux | OK | Mar 19 2025 |
Exports:AdPaik_1DAdPaikModelcoefsefrailty_sdfrailty_sd.AdPaikplot_bas_hazardplot_frailty_sdplot_ll_1Dplot_ll_1D.AdPaikplot_post_frailty_estplot_survivalsummarysurvival
Dependencies:
Citation
To cite package ‘TimeDepFrail’ in publications use:
Ragni A, Romani G, Masci C (2025). TimeDepFrail: Time Dependent Shared Frailty Cox Model. R package version 0.0.1, https://CRAN.R-project.org/package=TimeDepFrail.
Corresponding BibTeX entry:
@Manual{,
title = {TimeDepFrail: Time Dependent Shared Frailty Cox Model},
author = {Alessandra Ragni and Giulia Romani and Chiara Masci},
year = {2025},
note = {R package version 0.0.1},
url = {https://CRAN.R-project.org/package=TimeDepFrail},
}
Readme and manuals
TimeDepFrail: Time-Dependent Shared Frailty Cox Models in R
TimeDepFrail is the ultimate R package for fitting and analyzing Time-Dependent Shared Frailty Cox Models. These models extend the traditional Shared (Gamma) Frailty Cox Models by incorporating a time-dependent frailty component, making it a robust tool for studying how unexplained heterogeneity in data evolves over time.
This package implements the methods discussed in "Centre-Effect on Survival After Bone Marrow Transplantation: Application of Time-Dependent Frailty Models" by C.M. Wintrebert et al. (2004).
Installation
You can install the development version of the package from GitHub:
devtools::install_github("alessandragni/TimeDepFrail")
Dataset data_dropout
data_dropout
The data_dropout dataset is used to exemplify the package.
It tracks the academic progress of students enrolled in 2012 over three academic years (six semesters). This dataset aims to explore the factors leading to student dropout.
The dataset is composed of four variables:
-
Gender: Categorical covariate indicating gender (Male or Female). -
CFUP: Numeric covariate representing the standardized number of credits or CFUs (Credito Formativo Universitario) passed by the student in the first semester. -
time_to_event: The time (in semesters) when a student decides to drop out. A value greater than 6.0 means the student did not drop out during the follow-up period. -
group: Categorical variable representing the student's course of study, with 16 levels from CosA to CosP.
Students are followed for a maximum of 6 semesters (3 academic years), from the end of first semester until they drop out or the follow-up ends.
Model execution
To fit a Time-Dependent Shared Frailty model, the following elements are required:
- data as
data.frame, e.g.data_dropout -
time_axisvector: The time intervals for which the model is applied. For example, in thedata_dropoutdataset, no events occur in the first semester, so thetime_axisstarts at the end of the first semester (t = 1) and ends at the end of the third year (t = 6). -
categories_range_minandcategories_range_maxvectors: Provide minimum (categories_range_min) and maximum (categories_range_max) bounds for each parameter category to constrain the optimization. -
formulaobject: Specify the relationship between time-to-event, covariates, and group. For the clustering variable (group), it must be provided ascluster(group)in the formula.
Once these elements are prepared, you can call the desired model using the AdPaikModel() function.
For full examples, refer to the Examples/ReplicationCode.R script.
Additionally, for guidance on selecting model parameters such as time_axis, categories_range_min and categories_range_max,
we recommend base these choices on insights gained after fitting a Time-Unvarying Shared Frailty model.
You can find a relevant example in the Examples/ReplicationCode.R, in the Appendix D Section.
Analyzing results
Several built-in methods are available to analyze the results of the fitted model:
- Model Summaries:
summary(),coef(),coefse(),confint() - Baseline Hazard Step-Function:
plot_bas_hazard() - Frailty Standard Deviation/Variance:
frailty_sd()andplot_frailty_sd() - Posterior Frailty Estimates:
plot_post_frailty_est() - Conditional Survival Function:
survival()andplot_survival()
These methods provide insightful visualizations and summaries to help you interpret your model results effectively.
Furthermore, also a support function suitable for the choice of the range of parameters and analysis of the 1D log-likelihood is available, AdPaik_1D().
To be aware of
- The
AdPaikModelmodel is optimized for fast computation although estimating certain coefficients (e.g.,MaleversusFemale) may vary slightly in computational time. Note that changing the reference category (e.g., usingMaleas the baseline) alters the coefficient estimates but not the overall log-likelihood or model fit. Users should choose reference categories based on interpretability rather than performance.
Authors and maintainers of the code
Alessandra Ragni (alessandra.ragni@polimi.it), Giulia Romani (giulia.romani@mail.polimi.it), Chiara Masci (chiara.masci@polimi.it).
