TernTables: A Statistical Analysis and Table Generation Web Interface for Clinical and Biomedical Research.

Area of Science:

• Clinical research methodology
• Biostatistics
• Health informatics

Background:

• Clinical research dissemination faces challenges due to administrative hurdles and inconsistent methods.
• Standardized workflows for descriptive and univariate analyses are lacking, impacting efficiency and reproducibility.

Purpose of the Study:

• To introduce TernTables, an open-source application and R package designed to streamline clinical data reporting.
• To automate the generation of descriptive and univariate statistical results for clinical research.
• To enhance transparency and reproducibility in clinical data analysis.

Main Methods:

• Development of a web application and R package integrating protected health information (PHI) screening.
• Implementation of a rule-based decision tree for automated selection of statistical tests (frequency-based, parametric, non-parametric).
• Generation of publication-ready summary tables in Microsoft Word, including dynamic methods text and R code.

Main Results:

• Validation on a landmark clinical trial dataset showed concordance with established biostatistical methods for descriptive and univariate analyses.
• TernTables successfully automates the creation of formatted results from raw clinical data.
• The system ensures transparency and reproducibility through generated R code.

Conclusions:

• TernTables standardizes routine descriptive and univariate statistical workflows in clinical research.
• The platform democratizes access to rigorous statistical methods, reducing technical and financial barriers.
• It complements formal statistical consultation, allowing focus on complex analyses and study design.