Clinical Trial Statistics Demo (R)

This repository is a reproducible, simulated clinical-trial statistics example in R. Designed to show implementation of typical clinical trial analyses.

What this demo includes

Typical analyses in clinical trial statistics include:

• Baseline characteristics tables
• Continuous endpoint analysis with ANCOVA
• Binary endpoint analysis with logistic regression
• Time-to-event analysis with Kaplan-Meier curves and Cox regression
• Repeated-measures analysis for longitudinal outcomes
• Subgroup analyses with a simple forest plot
• Adverse event summaries

The data in this repo are simulated, so it is safe to run and share.

Files

• R/analysis.R — generates the dummy data, runs the analyses, and saves tables/figures
• results/ — CSV and text outputs created by the script
• figures/ — PNG figures created by the script

How to run

From the repository root:

Rscript R/analysis.R

After running, you should see outputs such as:

• results/baseline_summary.csv

Variable	Placebo	Active
Age, mean (SD)	58.4 (11.3)	59.2 (11.4)
Female, n (%)	50 (41.7%)	59 (49.2%)
High strata, n (%)	51 (42.5%)	52 (43.3%)

• results/continuous_endpoint.csv

Endpoint	Estimator	Estimate	CI_low	CI_high	P_value
Week 12 change from baseline	ANCOVA (Active vs Placebo)	-8.16	-10.5	-5.81	7.63e-11

• results/binary_endpoint.csv

Endpoint	Estimator	OR	CI_low	CI_high	P_value
Responder at week 12	Logistic regression OR (Active vs Placebo)	6.52	3.59	11.83	6.87e-10

• results/time_to_event.csv

Endpoint	Estimator	HR	CI_low	CI_high	P_value
Time to first event	Cox PH HR (Active vs Placebo)	0.59	0.42	0.81	0.00112

• results/mmrm_summary.txt

• results/subgroup_forest.csv

Subgroup	Level	Estimate	Low	High
strata	High	-9.27891183076301	-12.7920968602426	-5.76572680128337
strata	Low	-7.06562206506464	-10.3324723227326	-3.7987718073967
sex	Female	-7.79284266278517	-11.4304741680476	-4.15521115752276
sex	Male	-8.50758852792628	-11.67145662151	-5.34372043434259

• results/adverse_events_summary.csv

trt	subject_count	subjects_with_any_ae	events_total	subjects_with_serious_ae	percent_any_ae	percent_serious_ae
Placebo	120	65	100	6	54.2	5
Active	120	84	146	15	70	12.5

Figures

• Figure 1. Kaplan-Meier curve for time to first event by treatment arm.
• Figure 2. Mean biomarker trajectory over time with SD bars.
• Figure 3. Forest plot of subgroup treatment effects for the continuous endpoint.

Notes on the analyses

The code uses a fairly standard trial-statistics workflow:

• lm() for the primary continuous endpoint
• glm(..., family = binomial()) for the responder analysis
• survival::survfit() and survival::coxph() for time-to-event analyses
• nlme::lme() for a repeated-measures model