subpat - Time-to-event on subpopulations and subgroups - 2020 Shiny Contest Submission

subpat - Time-to-event on subpopulations and subgroups

Authors: Stefan Eng, Mustapha Larbaoui
Working with Shiny < 1 year

Abstract: Exploratory time-to-event analysis with subgroups and subpopulations via easy to use graphical interface. Supports Kaplan-Meier plots and Cox Ph models.

Full Description: subpat is a collection of package to help with common tasks for people in pharma doing exploratory data analysis using ADaM datasets. The main features include

  • Creating and editing subpopulations
  • Creating subgroups
  • Basic time-to-event analysis with flexible variable mapping
    • Includes Kaplan-Meier plots and Cox Ph models

Users such as medical writers can use the subpopulation features (via the PLG app in subpat) to create ad-hoc reports. Developers can easily plug-and-play the modules into their own applications.

One of the novel features of subpat is the use of tidymodules to manage complex nested modules in shiny. Tidymodules provides a new object-oriented programming (OOP) approach for module development, new module interface using input/output ports and a set of tidy operators for handling cross-module communication.

The bulk of this work was developed over the summer during my internship at Novartis in the SCC (Scientific computing and consulting) group.

Category: Healthcare
Keywords: pharma, bs4Dash, tidymodules
Shiny app:
RStudio Cloud:


Full image:

Subpopulations editing:

Subgroup creation:

KM Plot:

Cox PH:

tidymodule module visualization:

This looks quite good. Is it intended that the user can load their own data? I did not see how, within the Load data section. I'd like to use this, if I could load my own data.
Also after all the good work you put into this, it could use a brief description within the start page.
Good luck in the contest!

Hi @gbshiny. Thanks for taking a look at the app. You should be able to upload your own xpt file now (will add support for more files soon).