Clinical TLF Visualization Platform

A comprehensive, production-ready Shiny application for visualizing clinical trial data following CDISC SDTM standards. This application demonstrates advanced features including cross-filtering, drill-down navigation, faceting, bookmarking, and user preferences.

Features

Core Capabilities

Modular Architecture: Scalable, maintainable codebase with reusable modules
SDTM Compliance: Supports standard CDISC domains (DM, AE, LB, VS)
TLF Visualization: Tables, Listings, and Figures optimized for clinical data

Advanced Features

Cross-Filtering: Filters cascade across all modules automatically
Drill-Down Navigation: Click tables/plots to explore detailed data
Faceting: Dynamic grouping and small multiples for comparative analysis
Bookmarking: Save and share application state via URL
User Preferences: Customize appearance, behavior, and defaults
State Management: Centralized reactive state for consistency

Project Structure

clinical_tlf_app/
├── app.R                          # Main application file
├── global.R                       # Global environment setup
├── ARCHITECTURE.md                # Detailed architecture documentation
├── README.md                      # This file
│
├── config/
│   └── app_config.R              # Configuration settings
│
├── data/
│   ├── create_sdtm_data.R        # Data generation script
│   └── sdtm/                     # SDTM datasets (RDS/CSV)
│       ├── dm.rds
│       ├── ae.rds
│       ├── lb.rds
│       └── vs.rds
│
├── utils/
│   ├── state_management.R        # Reactive state management
│   ├── bookmark_utils.R          # URL bookmarking utilities
│   └── data_utils.R              # Data processing functions
│
├── modules/
│   ├── core/
│   │   ├── filter_module.R       # Cross-domain filtering
│   │   └── preferences_module.R  # User preferences
│   │
│   └── tlf/
│       ├── tlf_table_module.R    # Summary tables
│       ├── tlf_listing_module.R  # Data listings
│       └── tlf_figure_module.R   # Interactive figures
│
└── www/
    └── custom.css                # Custom styling

Installation

Prerequisites

R (>= 4.0.0)
RStudio (recommended)

Required Packages

install.packages(c(
  "shiny",           # Web application framework
  "bslib",           # Modern UI theming
  "DT",              # Interactive tables
  "plotly",          # Interactive plots
  "ggplot2",         # Static plots
  "dplyr",           # Data manipulation
  "tidyr",           # Data tidying
  "shinyWidgets",    # Enhanced widgets
  "jsonlite"         # JSON parsing
))

Setup

Clone or download the repository
Navigate to the clinical_tlf_app directory
Generate sample data:
```
source("data/create_sdtm_data.R")
```
Run the application:
```
shiny::runApp()
```

Usage Guide

Getting Started

Overview Tab: View study summary and quick statistics
Filters: Use the sidebar to apply filters across all domains
Tables: Generate summary tables grouped by various variables
Listings: Explore detailed record-level data
Figures: Create interactive visualizations with faceting

Cross-Filtering Workflow

Navigate to any tab (Tables, Listings, or Figures)
Open the filter sidebar
Select filters from any domain (DM, AE, LB, VS)
Click "Apply" to update all modules
Filters persist across tab navigation

Example:

Filter to "Drug 50mg" treatment arm
Filter to "SEVERE" adverse events
All tables, listings, and figures update automatically

Drill-Down Navigation

From Tables:

Select rows in summary tables
View automatically filters to selected subgroups

From Listings:

Click any row to see full record details
Detail panel shows all fields for the selected record

From Figures:

Click plot points to filter to specific subgroups
Use plotly tools to zoom, pan, and explore

Faceting

Navigate to the Figures tab
Enable "Faceting" checkbox
Select facet variable (ARM, SEX, SITEID)
Choose facet layout (Wrap or Grid)
Plot automatically splits into small multiples

Bookmarking

Apply desired filters and navigate to desired view
Click "Bookmark" button in header
Copy the generated URL
Share URL with colleagues or save for later
Opening the URL restores exact application state

User Preferences

Click "Preferences" button in header
Customize:
- Theme (light/dark)
- Default page length
- Default grouping variable
- Export settings
Click "Save Preferences"
Settings persist across sessions

Data Domains

DM (Demographics)

Subject-level baseline characteristics
Variables: USUBJID, AGE, SEX, RACE, ARM, SITEID
Used for: Demographics tables, baseline summaries

AE (Adverse Events)

Safety event data
Variables: AETERM, AESEV, AESER, AEREL, AEOUT
Used for: Safety tables, AE visualizations

LB (Laboratory)

Clinical laboratory results
Variables: LBTESTCD, LBSTRESN, LBNRIND, VISIT
Used for: Lab tables, time-series plots, shift tables

VS (Vital Signs)

Physiological measurements
Variables: VSTESTCD, VSSTRESN, VISIT
Used for: Vital sign trends, safety monitoring

Configuration

App Configuration (`config/app_config.R`)

Customize application behavior:

Data paths and formats
UI theme and styling
Feature flags (enable/disable features)
Domain-specific settings
Plot colors and defaults

Example:

app_config$ui$theme <- "darkly"  # Dark theme
app_config$features$enable_bookmarking <- FALSE  # Disable bookmarking

Adding New Domains

Create data generation function in create_sdtm_data.R
Add domain configuration in app_config.R
Update load_sdtm_data() in data_utils.R
Create module in modules/tlf/ (optional)
Add filters in filter_module.R

Advanced Usage

Custom Analysis

Add custom analyses by:

Creating new module in modules/
Sourcing module in global.R
Adding UI/server in app.R
Connecting to app_state for cross-filtering

Database Integration

Replace file-based data with database:

Modify load_sdtm_data() in data_utils.R
Use DBI package to connect to database
Implement lazy loading for large datasets
Add caching for performance

Export Enhancements

Extend export capabilities:

Add export buttons to modules
Use writexl for Excel export
Use rmarkdown for PDF reports
Include filter metadata in exports

Performance Considerations

Data Size: Tested with 150 subjects, ~15,000 records
Scalability: Use database backend for >100,000 records
Caching: Implement caching for expensive computations
Lazy Loading: Load modules on-demand
Async: Use promises for long operations

Development

Testing

Unit tests can be added using:

shinytest2 for module testing
testthat for utility function testing

Deployment

Deploy to:

Shiny Server (open source)
RStudio Connect (enterprise)
ShinyApps.io (cloud)
Docker (containerized)

Contributing

To extend the application:

Follow modular architecture patterns
Use state management for cross-module communication
Document new features in ARCHITECTURE.md
Add configuration options in app_config.R

Troubleshooting

Common Issues

Application won't start:

Check all packages are installed
Run source("global.R") to check for errors
Verify data files exist in data/sdtm/

Filters not working:

Ensure filter module is connected to app_state
Check get_filtered_data() is called in modules
Verify trigger_update is incrementing

Bookmarks not restoring:

Check URL parameters are present
Verify restore_from_url() is called
Ensure enableBookmarking() is set

License

This is a demonstration project for portfolio/professional purposes.

Contact

For questions or issues, please refer to the project documentation or contact the development team.

Version: 1.0.0 Last Updated: 2025-10-20 Built with: R Shiny, bslib, plotly, DT

Name		Name	Last commit message	Last commit date
config		config
data		data
modules		modules
utils		utils
www		www
ARCHITECTURE.md		ARCHITECTURE.md
README.md		README.md
USAGE_GUIDE.md		USAGE_GUIDE.md
app.R		app.R
global.R		global.R

Latest commit

History

Folders and files

Repository files navigation