AI-Enhanced Clinical Trial Design for Hidradenitis Suppurativa
In healthcare and pharmaceutical research, clinical trial success often depends on selecting the right patient population from the start. This case study shows how Langate helped a global pharmaceutical company improve the design of its hidradenitis suppurativa trial by using AI to uncover predictive patterns in historical clinical data.
Summary
Langate implemented an AI-powered trial optimization platform that analyzed historical dermatology studies to improve patient selection and study planning.
Quick Facts
- Timeline: 4 months
- Team: 6 specialists
- Output: 4 predictive baseline variables identified
- 40%+ improvement in predicted trial success probability
- Trial setup timelines have been significantly reduced
Project Scope
The project included integrating clinical datasets, developing machine learning models, and deploying visualization tools for research and clinical teams. The solution was designed to support future trial-planning workflows within a secure, compliant environment.
Tech Stack
- Python
- Scikit-learn
- Vertex AI
- Spotfire
Data sources: ClinicalTrials.gov, Pharmaproject, PubMed, Internal Trial Data
The Client
The client is a global pharmaceutical company conducting multiple dermatology clinical trials. Their objective was to improve trial design for hidradenitis suppurativa by identifying patient characteristics most strongly associated with successful treatment outcomes.
Business Requirements
- Improve clinical trial design and patient selection
- Reduce risk of unsuccessful trial outcomes
- Accelerate trial setup and planning
- Support evidence-based enrollment criteria decisions
- Create a repeatable framework for future studies
The Problem
Clinical outcomes in HS trials varied significantly, making it difficult to define optimal enrollment criteria. Manual review of historical data was time-consuming and inconsistent, limiting the ability to identify subtle but clinically meaningful correlations.
The Challenges
-
Working with large volumes of historical trial and clinical data - Ensuring reliability and interpretability of AI-generated insights
- Integrating predictive analytics into existing trial planning workflows
- Maintaining compliance and data security requirements
Problems Revealed
The project revealed that many predictive patient characteristics were hidden within complex combinations of variables that traditional review methods could not easily detect. This limited the organization’s ability to consistently optimize trial enrollment strategies.
The Solution
Achievements
-
Identified 4 baseline variables strongly associated with positive outcomes
-
Increased predicted trial success probability by more than 40%
-
Reduced study planning timelines from months to weeks
-
Improved confidence and consistency in trial design decisions
The platform enabled the client to shift from experience-based planning to a more predictive, data-driven clinical trial strategy.
Future Plans
-
The client plans to extend the platform across additional dermatology and immunology programs.
-
Langate continues to support the development of AI-driven capabilities for adaptive trial design and predictive patient stratification.
Ready to Transform Clinical Trial Design With AI?
Send a message to Langate sales
