Modeling Overview

ML • MMRM • PK/PD • Survival — reproducible, review-ready analyses

TipGoal

This section showcases analysis-grade modeling workflows with clear assumptions, validation, and traceability.

Model → Analysis Dataset → SDTM → Source

Tracks

Machine Learning (ML)

Interpretable predictive modeling + leakage-safe validation.

MMRM (Repeated Measures)

Longitudinal continuous endpoints using mixed models.

  • Baseline adjustment + visit effects
  • Covariance structure + inference
  • LS-means + contrasts
  • Go to MMRM

PK/PD Modeling

Exposure–response and concentration-time modeling.

  • Structural + error models
  • Covariates + diagnostics
  • Simulation-ready reporting
  • Go to PK/PD

Survival / Time-to-Event

KM/Cox and modern survival workflows.

  • KM + medians + at-risk
  • Cox HR + diagnostics
  • Sensitivity checks
  • Go to Survival

What you’ll find here

  • Purpose: Practical modeling examples aligned with clinical analysis expectations
  • Traceability: Model outputs link back to analysis datasets, specs, and programs
  • Review readiness: Assumptions stated, QC checks documented, results reproducible