R Clinical Analysis Workflow

Regulatory-compliant clinical trial analysis workflow. This command orchestrates the biostatistician agent with support from viz-specialist and reporting-engineer for comprehensive clinical data analysis.

Workflow Steps

1. Data Validation

• Verify CDISC-compliant data structure (ADaM/SDTM)
• Check required variables for analysis type
• Validate data integrity and completeness

2. Population Definition

• Define analysis populations (ITT, mITT, PP, Safety)
• Generate disposition summary
• Create CONSORT diagram

3. Statistical Analysis

Perform analysis based on endpoint type:

• Continuous endpoints: MMRM, ANCOVA, change from baseline
• Binary endpoints: CMH, logistic regression, responder analysis
• Time-to-event: Kaplan-Meier, Cox regression, RMST
• Count data: Poisson/negative binomial models

4. Tables, Listings, and Figures (TLF)

Generate standard clinical outputs:

• Demographic tables (Table 14.1)
• Efficacy tables
• Safety tables (adverse events, labs)
• Forest plots for subgroups
• Kaplan-Meier curves

5. Report Generation

Compile analysis into regulatory-ready format:

• Statistical Analysis Report (SAR)
• Define file outputs
• Include validation documentation

Usage

/r-clinical-analysis [data_path] [analysis_type] [output_format]

Parameters

• data_path: Path to ADaM dataset(s) or directory
• analysis_type: One of efficacy, safety, survival, pk, full
• output_format: One of html, pdf, rtf, docx (default: html)

Analysis Types

efficacy: Primary and secondary efficacy endpoints

• MMRM for repeated measures
• ANCOVA for single timepoints
• Responder analysis

safety: Safety analysis

• Adverse event summary
• Laboratory parameters
• Vital signs
• Exposure summary

survival: Time-to-event analysis

• Kaplan-Meier estimation
• Log-rank test
• Cox proportional hazards
• Landmark analysis

pk: Pharmacokinetic analysis

• Summary statistics
• Concentration-time profiles
• Non-compartmental analysis

full: Complete clinical study report analysis

Output Location

output/
├── code/
│   └── clinical_analysis.R
├── reports/
│   └── statistical_analysis_report.html
├── tables/
│   ├── t_dm.rtf                # Demographics
│   ├── t_efficacy.rtf          # Efficacy
│   └── t_ae.rtf                # Adverse events
├── figures/
│   ├── f_km.png                # Kaplan-Meier
│   └── f_forest.png            # Forest plot
└── listings/
    └── l_ae.rtf                # AE listing

Example

/r-clinical-analysis data/adsl.sas7bdat,data/adae.sas7bdat,data/adtte.sas7bdat full rtf

This performs complete clinical analysis and generates RTF outputs for regulatory submission.

Regulatory Compliance

ICH E9 Compliance

• Estimand specification
• Intercurrent event handling strategies
• Sensitivity analyses

21 CFR Part 11

• Audit trail documentation
• Version control
• Validation documentation

CDISC Standards

• ADaM variable naming
• Standard analysis flags
• Define.xml compatibility

Report Contents

1. Executive Summary

   • Key efficacy results
   • Safety summary
   • Conclusions

2. Methods

   • Analysis populations
   • Statistical methods
   • Handling of missing data

3. Results

   • Efficacy analyses
   • Safety analyses
   • Subgroup analyses

4. Appendices

   • Detailed tables
   • Statistical code
   • Validation documentation

Packages Used

• survival, survminer: Survival analysis
• mmrm: Mixed models for repeated measures
• emmeans: Estimated marginal means
• gtsummary, gt: Clinical tables
• rtables, tern: Pharmaverse tables
• haven: SAS data import

Notes

• All analyses follow pre-specified SAP
• Original data is never modified
• Validation documentation is generated
• Code is audit-ready with comments