Pharmacovigilance

Adverse event data extraction and coding from unstructured sources accelerates drug safety monitoring and reporting

Business impact

• Time to identify safety issues — Reduces delay in spotting adverse drug reactions by automating data extraction
• Pharmacovigilance case processing efficiency — Increases throughput by minimizing manual review and data entry tasks
• Regulatory reporting speed — Speeds up submission of safety reports to regulators through automation

Data requirements

• Unstructured and semi-structured source documents (Text) — Extract adverse event information from diverse text formats using AI
• Patient support program records (Text) — Analyze real-world patient data to identify safety signals
• Social media and online forums (Text) — Mine patient-reported experiences for adverse drug event detection
• Call center voice recordings (Audio) — Transcribe and analyze audio data to detect potential adverse events

AI methods and techniques

• Predictive AI — Forecast potential adverse events and prioritize cases for review
• Generative AI — Generate regulatory-compliant case summaries and narratives automatically
• Symbolic AI — Use knowledge graphs to ground AI in pharmacological domain expertise

AI models and model families

GPT-4o, Gemini, T5, BERT

Industries

Real-world evidence

2 documented case studies on record.

Companies using this: Bayer, medac GmbH.