CogStack

CogStack enables efficient retrieval, processing, extraction, and de-identification of structured and unstructured clinical data to support clinical service improvement and research.

Key Features:

• Distributed Architecture: A lightweight distributed database processing architecture that provides fault tolerance and scalability for large-scale clinical data processing.
• Clinical data processing: Handles retrieval, processing, and extraction from both structured and unstructured clinical data sources, including patient records and clinical notes.
• Natural Language Processing (NLP): Performs NLP preprocessing and processing to extract insights from clinical text.
• De-identification algorithm: Implements a de-identification algorithm reported at up to 94% precision and 96% recall on simulated clinical text.
• Information governance compliance: Supports adherence to NHS information governance policies through controlled processing and de-identification of clinical data.

Scientific Applications:

• Large-scale clinical data processing: Deployed at King's College Hospital to process over 300 million lines of clinical data for service and analytics purposes.
• Clinical research and service improvement: Provides timely access to relevant clinical information for clinicians, researchers, and business intelligence stakeholders.
• Genomic data integration: Supports initiatives to integrate genomic data into clinical practice.

Methodology:

Uses a lightweight distributed database processing architecture, NLP preprocessing and processing of clinical text, and a de-identification algorithm (reported up to 94% precision and 96% recall on simulated clinical text).