Overview

Large amounts of clinical data is locked up in free-hand notes, and other document formats that were not originally designed for entry into computer systems. How can this data be extracted for the purposes of standardization, consolidation, and, ultimately, for clinical research? LabKey Server's Natural Language Processing Pipeline (NLP) and document abstraction workflow tools help to unlock this data and transform it into a format that can better yield clinical insight.

LabKey Server's solution focuses on the overall workflow required to efficiently transform large of amounts of data into formats usable by researchers. Whether your group is already using an NLP engine, developing algorithms to support one, or doing all document abstraction manually, the process requires the following:

  • A human and/or machine process that prepares the documents for submission to the process.
  • Fault-tolerant integration of multiple independent systems, so that your workflow can recover from interruptions in service.
  • A system of human and/or machine abstraction that applies algorithms and knowledge to abstract the data from the documents.
  • Scoring and review of results at multiple points within the workflow.
  • Feedback on algorithms or human abstractors to improve reliability.
  • Flexible scalability as new scenarios and data sources are added to the overall workflow.
The LabKey NLP workflow system allows teams to take an integrated, scalable approach to both manual data abstraction and automated natural language processing (NLP) engine use. The system enables teams to efficiently and reproducibly transform clinical data into formats useful for both research and care.

LabKey NLP is available for subscription purchase from LabKey. For further information, please contact LabKey.

Documentation

Resources

discussion

Discussion

Search 

Pages 

previousnext
 
expand all collapse all