Classifying Depression Severity with Real-World Data
NLP-powered insights into cognitively impaired schizophrenia (CIAS) patients in UK secondary care
Partner: Global Pharma Company
CHALLENGE
- No reliable way to classify depression severity from routine data
- Mild cases often missing, so bias is skewed towards severe patients
OUR APPROACH
- Leveraged Akrivia’s RDLA platform and clinical input
- Build and validated rule-based classifier (asymptomatic → severe)
OUTCOMES
- Clear patient cohort and severity spectrum
- Deeper insight into depression trajectories
- Strong base for advanced model (LLMs)
IMPACT
- Improved patient stratification
- Faster, lower risk drug development
“Working with your team was exceptional. Transparent, open, and truly collaborative from start to finish. Project management was spot-on—clear timelines, smooth delivery, and remarkable flexibility throughout.
While results can’t be guaranteed in projects like this, you made the process seamless and set us up perfectly to advance our UK linked database work. Grateful for the partnership—can’t wait to collaborate again.”
– Senior Manager, HEOR, Compass Pathways