Hua Xu, Jon Duke, George Hripcsak, Karthik Natarajan, Anupama Gururaj, Min Jiang, Alexandre Yahi, Noemie Elhadad, Juan M Banda, Olga Patterson, Scott DuVall, Vojtech Huser

nlp_wg_meeting_03092016.pdf

1. Note-type mapping Presentation – Karthik
2. Share existing ontologies from Vanderbilt (Hua) and Regenstrief (Jon)
3. Share strategies for combining data from different searches – Jon
4. Report on WG for commenting – Hua
5. Wrappers for cTAKES and Metamap – Min
6. Improvements to search engine set up using MT samples – Min
7. Textual Data Representation – Discussion
8. Goals of 2016
9. Change of meeting time

1. Note-type mapping Presentation – Karthik
   1. the model is based on the SHARE-N model and adapted to the current data structure. This model incorporates other semantic types and all of the modifiers are not available in cTAKES yet.
   2. the notes were processed from eMERGE cohort at Columbia with about 60,000 notes encompassing 1700 patients. The original patient number was 3200.
   3. In theory, a set containing the combination of minimal modifiers can be generated. Practically, can we trust the data enough to add it into OHDSI tables? - only highest confidence data (with maximum PPV) should be added to the tables.
   4. Next steps:
      1. Look at the note sections to determine the errors.
      2. Work with Sunny to generate the NLP outputs for the phenotyping data
      3. Evaluate by comparisons with structured data
      4. Make the system more robust
      5. Generate a protocol and/or annotation guidelines
      6. Share the data as a Gold standard with manually annotated CUIs
      7. Alex's script is to be tried on different datasets and evaluated across notes from different institutions
      8. Identify minimal set of notes to work with when recommending to the OHDSI community
      9. Identify sets of concepts that are not reliable - negation is a very good example of this idea.
      10. Continue discussion of NLP system evaluation across different sites
2. existing ontologies from Vanderbilt - Hua - document-types.zip

1. Share existing ontologies from Regenstrief (Jon)
2. Share strategies for combining data from different searches - Jon
3. Report on WG for commenting - Hua
4. Wrappers for cTAKES and Metamap - Min
5. Improvements to search engine set up using MT samples - Min
6. Textual Data Representation - Discussion
7. NLP system evaluation across different sites - Discussion