User Tools

Site Tools


Meeting Notes August 29, 2019


  • Robert Miller asked about how code that prepares data for the common ETL input format should deal with ambiguous ICDO site/histology combination in multiple NAACCR schemas.
  • It was agreed that the task 'Create ICDO concept for 'Unknown Histology' and precoordinate with all ICDO sites.' is not on the critical path for the symposium.
  • We reviewed the vocabulary tasks necessary for the symposium:
    • Fix missing leading zeros in the concept_codes for Treatment NAACCR item codes.
      • It was discussed that this task is only confined to treatment variables. But that other leading zero problems might exist as well for other NAACCR variables. Dymtry will manually fix the NAACCR treatment variables.
    • The 'NAACCR Proc Schema' entries do not have all the necessary relationships to ICDO codes via 'Proc Schema to ICDO'.
    • Add 'Metatdata' 'Table' and 'Field' entries for EPISODE and EPISODE_EVENT and new columns in MEASUREMENT.
      • Dymytry said that the building of these Metadata concepts is part of the vocab release process. Since the DDL for the extension tables is not part of an official release, these concepts don't get created. For the time being, Dymytry will provide script in the OncologyWg repository that ETLers will need to run.
    • Move NAACCR Treatment 'did not happen' concepts into the Observation domain.
      • It was agreed that for the time being only Treatment NAACCR item codes describing flavors of 'did not happen' or 'unknown' will be moved into the Observation domain. Diagnosis and Treatment modifiers could possibly in the future also be placed in the Observation table. Jermey Warner raised the possibility of adding a 'Plan' domain to OMOP to represent these kinds expectational/intentional data.
    • Create new 'Regimen' domain, move all the regimens into the new domain, create entries in the new domain for NAACCR high-level drug concepts and non-standardize existing drug NAACCR item codes.
      • It was agreed to move 'Regimen Type' classification concepts into the new 'Regimen' domain. It was recognized that 'Regimen Type' are exact matches for new high-level Drug NAACCR concepts and that this is OK for the time being.
    • Disambiguate the “OR” regimens into multiple regimens.
  • We discussed timelines for Symposium.
    • Dymytry said that the vocabulary will aim to have the tasks done by 9/5
    • The group agreed that analysis code should aim to be done by the 9/9
    • The group agreed that ETL work shold aim to be done by 9/9.
  • We discussed Andrew Williams concerns about how to frame the results.
    • The group agreed that we should frame the results as a beta test proof-of-concept. That many problems, flaws and challenges exist but that have a version 1 accomplished.
documentation/oncology/meeting_notes_2019_aug-29.txt · Last modified: 2019/08/30 13:50 by mgurley