 |
Information Extraction with TeSSI® Extraction Engine
Usually when healthcare workers are reviewing patient data in free text documents, they
are looking for very specific and discrete pieces of
information such as problems,
allergies, or current medications. This seemingly simple task is not easy as it should be
due to the following characteristics of typical free text documentation:
- They often contain superfluous information that is not relevant to the knowledge the reader
currently needs;
- They are often quite long with the desired information spread throughout the document;
- Differences in style among doctors can lead to inconsistent
structure and granularity of
details in the document;
- Their are multiple documents in which the desired information is located.
All of these issues can be addressed with an
automated and "domain-knowledgable" Information Extraction (IE) tool.
L&C’s IE technology automatically
extracts
discrete data from free text to a format dictated by the application. Using both syntactic and semantic information,
IE can "fill-in the blanks" of pre-configured templates with
the required relevant information hidden inside of free text documents.
original free text
discharge summary |
resulting template with
extracted information |
TeSSi Indexing
TeSSi Search
|
|
