profile

Smart Data Analysis and Statistics

We Trained an AI on Evidence Synthesis. Want to See What It Can Do?

Published 9 days ago • 2 min read

​

Dear Reader:

​

Are you GenAI-curious?

With all the buzz about how AI is set to transform pharma—and just about every other industry—it’s hard not to be intrigued. Maybe you’ve already started experimenting with tools like ChatGPT. As a technologist and AI expert, I’ve gone deep down that path myself. And here’s what I’m seeing:

Large Language Models (LLMs) are everywhere—and quickly becoming commoditized. The real differentiator won’t be the model itself, but how it’s trained, validated, and implemented. That’s why I developed the SDAS AI Assistant—a purpose-built tool designed to tackle real-world challenges in pharma and research.

I’m now opening up beta testing, and I’d love for you to be part of it.

"SDAS AI Advisor is the next best thing to contracting Thomas Debray 24/7."

​

"Interesting! so what does being a beta tester involve?"

​

Probably about a half hour, maybe less, to run your current draft or finalized content through SDAS AI Assistant to see how it performs, and provide us with feedback.

The current use case for the SDAS AI Assistant is to review draft or finalized reports of indirect treatment comparisons (ITCs) and network meta-analyses (NMAs). Acting as a proxy for its trainer, Thomas Debray, PhD, the Assistant evaluates the conduct and reporting of the analysis and offers recommendations for improvement.

Specifically, SDAS AI Assistant will:

  • Check compliance with guidelines for the conduct and reporting of ITC/NMA
  • Assess the research question and objectives
  • Review the methodology and determine if it is sufficiently described
  • Evaluate whether assumptions are appropriate and well-justified
  • Assess the scope and adequacy of sensitivity analyses performed
  • Identify gaps or limitations and suggest improvements
  • And more!

The feedback we are looking for is:

  • Was the output useful, relevant, and productive?
  • What could make it better?
  • Any other suggestions or feedback.

The genAI revolution is a tide that will float all boats. Or sink some. Let's collaborate today! Have a report you're either working on, or curious how it would fare against a highly trained, finely tunned LLM? Let's discuss, book a quick call today: here​

Sincerely,

Thomas Debray, PhD
Founder & Owner
Smart Data Analysis and Statistics B.V.

​tdebray@fromdatatowisdom.com​

Let's talk evidence synthesis, RWE and AI - Book some time here​

SDAS Industry Events: 2025

We are attending several industry events in the coming months, if you or someone on your team is attending as well, let's connect and compare notes- see below!

​

Handbook of Comparative Effectiveness and Personalized Medicine using RWE (2025)

Edited by Thomas Debray, R.W. Platt & L. Nguyen

A practical guide to estimating treatment effectiveness in real-world populations. Learn how to leverage real-world data (RWD) — alone or combined with other sources—to generate both overall and individualized treatment effect estimates, and more!

SimTOST

Revolutionizing Equivalence Trial Design

​

Optimize sample size estimation for randomized bioequivalence trials with powerful simulation capabilities.

  • Supports multiple correlated primary endpoints
  • Handles multiple treatment arms
  • Built-in multiplicity corrections
  • Optimized with C++ for fast simulations
  • Extensive vignettes

SimTOST is open-source, combining the flexibility of R with the speed of C++ for computational efficiency. The package has recently been released on CRAN.

AI-Enhanced NMA: Smarter, Faster, HTA-Compliant

Outsourcing evidence synthesis can mean high costs, long timelines, and less control. Our AI-powered tool streamlines Network Meta-Analysis, delivering HTA-compliant reports efficiently—giving you control without compromising quality.

Make smarter clinical decisions, faster.
​

Smart Data Analysis and Statistics

Thomas Debray, PhD

Stay ahead with our newsletter, crafted for detail-oriented statisticians committed to making a meaningful impact. Discover advanced methods for clinical trial and real-world evidence studies, strategies to enhance transparency and reproducibility, and updates on essential trainings and workshops.

Read more from Smart Data Analysis and Statistics
Smart Data Analysis and SimTOST R package for bioequivalence trials.

Dear Reader: Welcome to the SDAS community! I’m excited to have you on board and look forward to sharing valuable insights, tools, and resources to support your work in data science and statistics. A bit about me I’ve had the opportunity to work at the forefront of data science and statistics in the pharmaceutical industry, where things are moving at light speed. My background includes: Assistant Professor at the University Medical Center, Utrecht Lead or contributing author to hundreds of...

about 1 month ago • 1 min read