Clinical Trials Statistical Analysis in R

Overview

This repository showcases advanced statistical analysis techniques applied to clinical trials data using R. It contains R scripts that focus on non-parametric statistical tests, permutation tests, and data visualization. The analysis includes clinical trial treatment comparisons, blood protein concentration studies, and rank-based hypothesis testing.

Files Included

1. ClinicalTrials_nonpar.R

• Purpose: Performs statistical analysis on clinical trial data using non-parametric methods.
• Key Features:
  • Implements Wilcoxon-Mann-Whitney (WMW) test for rank-sum comparisons.
  • Uses permutation tests to approximate the null distribution.
  • Conducts Kruskal-Wallis tests for multi-group comparisons.
  • Analyzes the effects of treatments on rat body weight and protein concentration.
  • Generates rank-based relative effect estimates.
• Functions Used:
  • wilcox.test(), kruskal.test()
  • rank(), table(), mean(), sd()
  • ggplot2 for visualization

2. ClinicalTrials_mctp.R

• Purpose: Provides statistical solutions for comparing treatment groups in clinical trials.
• Key Features:
  • Computes permutation-based probability mass functions (PMF).
  • Simulates the null distribution of rank-based test statistics.
  • Generates confidence intervals for relative effects.
  • Implements multiple contrast test procedures (MCTP) to compare treatment conditions.
• Functions Used:
  • sample(), sum(), mean(), quantile()
  • pander(), ggplot2, boxplot(), stripchart()
  • npar.t.test() from the nparcomp package for relative effects

Features

• Statistical Analysis
  • Wilcoxon-Mann-Whitney test, Kruskal-Wallis test
  • Non-parametric multiple contrast test procedures (MCTP)
  • Rank-based confidence intervals
• Visualization
  • Boxplots, density plots, and histograms (ggplot2)
  • Permutation distributions and rank-based PMF plots
• Reproducible Reports
  • Well-structured code for easy replication of results
  • Uses pander and knitr for formatted output

Installation & Requirements

Ensure you have the required R packages installed:

install.packages(c("tidyverse", "ggplot2", "dplyr", "nparcomp", "multcomp", "pander"))

Usage

1. Clone the repository:

   git clone https://github.com/yourusername/Clinical-Trials-Stats-R.git
   cd Clinical-Trials-Stats-R

2. Run the scripts in RStudio or R console:

   source("ClinicalTrials_nonpar.R")
   source("ClinicalTrials_mctp.R")

Author

Abhinav Mishra
Master's in Bioinformatics, FU Berlin

License

MIT License