search

National-Clinical-Cohort-Collaborative / Data-Ingestion-and-Harmonization

Data Ingestion and Harmonization
41 stars 12 forks source link

Issue raised by PamBanning regarding the LOINC code correction via TestNorm tool #70

Closed stephanieshong closed 2 years ago

stephanieshong commented 4 years ago

PamBanning 2:53 PM Hi Stephanie. This looks pretty straightforward from the displays, and I understand the goal is not to be as granular as possible. I have a lot of apprehension about COVID-19 displays going to PCR, as many can be antigen or antibody, but finally realized this week that antigens, antibodies and nucleic acids were being grouped as one in this project. I wonder about the Specimen Source (row 2). Might having that go to the SARS CoV 2 RNA assay artificially inflate the numbers of PCR testing? Usually any Specimen Source field is coded to 31208-2 Specimen source [Identifier] of Unspecified specimen . Would that make it too granular, where the idea is to have it attached as supportive information to the PCR test?

stephanieshong commented 2 years ago

FOR PCORNet CDM if the COVID19 test's LOINC code is null then we use the testnorm mapping table to map the LOINC code based on the COVID19LabTestNames. The curated dataset is here: https://unite.nih.gov/workspace/data-integration/dataset/preview/ri.foundry.main.dataset.c3aebef7-8250-47f0-be8b-d0f1fb46d17a/master. The comparison is done case insensitive to match on all case variations.