QC-TLF Python Verification

Project Overview

This project generates deterministic, reproducible QC Tables, Listings, and Figures (TLFs) from CDISC SDTM and ADaM datasets for independent verification against SAS-generated TLFs. It is designed for cell-by-cell comparison in clinical trial programming QC, not for regulatory submission.

Key Features

• Input: SDTM and ADaM datasets (CSV/XPT)
• Output: QC TLFs (CSV, Excel, PNG)
• Strict SAP/SAS logic adherence
• Tools: pandas, numpy, matplotlib
• Deterministic, reproducible results
• Explicit documentation of dataset/variable names, population flags, analysis parameters, grouping, ordering
• Handles missing values, sorting, rounding, decimal precision

Project Structure

1. Load Data
2. Apply Population Filters
3. Apply Derivation Logic
4. Generate TLFs
5. Export Output
6. Generate Listings

Sample Outputs

• qc_outputs/table1_age_by_treatment.csv: Summary statistics for AGE by treatment arm
• qc_outputs/listing1_subject_safety.csv: Subject-level listing for Safety population
• qc_outputs/figure1_age_histogram.png: Histogram of AGE by treatment arm

Setup Instructions

1. Clone the repository:

   git clone https://github.com/justin-mbca/QC-TLF-Python-Verification.git

2. Install Python dependencies:

   pip install pandas numpy matplotlib

3. Place your SDTM and ADaM datasets in the appropriate folders (sdtm/, adam/).
4. Run the main script:

   python TLF.py

Usage

• Modify the scripts as needed to match your SAP and SAS logic.
• Outputs will be generated in the qc_outputs/ folder.
• Compare outputs with SAS-generated TLFs for QC.

Calling SAS from Python

• Use Python’s subprocess module for local SAS execution.
• Use saspy for remote SAS server or SAS Viya integration.
• See run_sas_from_python.py for examples.

Marp Slides

• Project slides are available in markdown files and exported PPTX.
• See QC_TLF_Project_Marp_Slides_Detailed.md and QC_TLF_Project_Marp_Slides_Detailed.pptx.

Contact

For questions or contributions, please open an issue or contact the repository owner.