T-Library

T-Library extracts clinical data from unstructured electronic medical records (EMRs) using text mining and deep learning to enable clinical research and patient risk assessment.

Key Features:

• Automated data extraction: Extracts critical clinical variables from unstructured EMR text.
• Text mining: Applies text mining techniques to interpret free-text clinical notes and records.
• Deep learning-based patient clustering: Uses deep learning algorithms to cluster patients by morbid states.
• Continuously updated risk engine: Generates a risk engine that predicts complications and is continuously updated from clustered patient data.

Scientific Applications:

• Clinical research data generation: Produces structured datasets from EMRs for use in clinical studies and analyses.
• Patient stratification and risk assessment: Enables clustering-based stratification and prediction of patient complications for risk modeling.
• Personalized medicine: Supports development of targeted treatment strategies by identifying patient subgroups based on morbid states.

Methodology:

Combines text mining techniques with deep learning algorithms to extract information from unstructured EMR text, cluster patients by morbid states, and generate a continuously updated risk engine.