Statistical Modeling
This section presents a practical, portfolio-oriented view of statistical modeling in clinical and biomedical workflows, with emphasis on scientific fit, assumption awareness, interpretability, and implementation readiness.
Overview
Modeling plays a central role in translating clinical questions into quantitative evidence. In this portfolio, I use modeling to support:
- Longitudinal outcome analysis (e.g., repeated measures)
- Time-to-event analysis (e.g., survival endpoints)
- PK/PD and exposure-response thinking
- Traceable, review-ready analytical workflows
My focus is not only on selecting the right method, but also on ensuring that methods are implemented in a way that is:
- scientifically appropriate
- transparent
- reproducible
- QC-aware
- practical for delivery
What this section covers
Modeling Methods (hub)
A method-focused overview of the main modeling families included in this portfolio, including how they connect to study objectives, assumptions, and outputs.
Modeling Overview
A broader framing of modeling workflows, including method selection, implementation considerations, and interpretation in a regulated analysis context.
MMRM
Repeated-measures modeling for longitudinal endpoints, including practical considerations for estimation, covariance structure, and reporting.
Survival Analysis
Time-to-event methods for clinical endpoints, including Kaplan-Meier workflows, event/censoring considerations, and inferential interpretation.
PK/PD Modeling
PK/PD and exposure-response modeling concepts and implementation framing, with attention to model purpose, assumptions, and communication of results.
How I approach modeling work
Question-driven method selection
I start with the endpoint, estimand/analysis objective, and decision context before choosing a modeling approach.
Assumption-aware implementation
I treat assumptions as part of the analysis deliverable—not hidden details. This supports better interpretation and more credible conclusions.
Traceability and reproducibility
I aim to make modeling workflows easy to follow from: - analysis objective - data inputs - modeling choices - outputs and interpretation
Delivery readiness
Modeling outputs are most useful when they are organized for review, aligned with study reporting needs, and supported by clear documentation.
Notes for portfolio reviewers
This section is designed to show how I combine statistical methodology with implementation discipline and review-ready communication in applied modeling workflows.