Snomed and LOINC - Systematic content validation

[jamlung-ri]

Joe to work with Jon to write up the basic requirements for scripts to validate the LOINC and SNOMED content that has been test-loaded into OCL Staging.

[jamlung-ri]

LOINC was validated automatically, but the SNOMED export is still processing. Need to push this until we can get that export.

[jamlung-ri]

I have been using Burke's validator script to automatically validate SNOMED to determine if all expected concepts and mappings are present. I have also been looking in the SNOMED hierarchy to look for unexpected results.

Here are my findings so far:

• There are 608 duplicate SNOMED concepts in OCL, although none of these seem to be duplicates in the import file. Examples: 104496000 and 86122002
• There are 5 concepts that are missing from OCL. However, when I attempt to reload the concept into OCL QA to test the syntax, it seems to work fine. Missing concept IDs include 255135009, 419602005, 115119004, 365951004, 365055005
• There is a large number of concepts where the hierarchy path (i.e. the chain of parent concepts up to the SNOMED Root concept) is broken or fragmented.
  • My first thought was that the import file might have listed these concepts out of order, so the hierarchy could not have been created correctly, but I did a test that showed that the sorting of the concepts was correct.
  • Examples include 441742003 and 102316004. If you look at these concepts, they will not have a parent concept, although they do have parents in the import file.

@paynejd This seems pretty significant. I will put this on the Arch agenda for this week so we can come up with a game plan.

[jamlung-ri]

Sunny, Jon, and I agreed that this many issues warrants the full deletion of SNOMED from OCL Staging. It will then be reimported and checked again for correctness.

[jamlung-ri]

LOINC Validation output:

{ "missing_concepts":[], "extra_concepts": [], "Concepts compared": 152184, "missing_mappings": [ {"type": "Mapping", "map_type": "Associated Observations", "to_concept_url": "/orgs/Regenstrief-Institute/sources/LOINC-3/concepts/81323-8/", "from_concept_url": "/orgs/Regenstrief-Institute/sources/LOINC-3/concepts/81323-8/", "source": "LOINC-3", "owner_type": "Organization", "owner": "Regenstrief-Institute"} ], "Mappings compared": 261092, "extra_mappings": [] }

Note that the missing mapping has since been added.

[jamlung-ri]

SNOMED Validation output:

{ "missing_concepts":[ {"type": "Concept", "id": "255135009", "concept_class": "none", "datatype": "string", "extras": {"NHS Clinical Terms Version 3": "X78lh", "UMLS CUI": "C0347084", "UMLS Semantic Type": "Neoplastic Process", "Concept Status": "Inactive", "MODULEID": "900000000000207008", "Code in Source": "255135009"}, "retired": true, "source": "SNOMED", "names": [{"locale": "en", "locale_preferred": true, "name": "Carcinoma in situ of labial mucosa (disorder)", "name_type": "Fully Specified"}, {"locale": "en", "locale_preferred": false, "name": "Carcinoma in situ of oral aspect of lip", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Carcinoma in situ of inner aspect of lip", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Carcinoma in situ of mucosa of lip", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Carcinoma in situ of buccal aspect of lip", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Carcinoma in situ of labial mucosa", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Carcinoma in situ of labial mucosa (disorder)", "name_type": "Synonym"}]}, {"type": "Concept", "id": "419602005", "concept_class": "none", "datatype": "string", "extras": {"NHS Clinical Terms Version 3": "XUdJK", "UMLS Semantic Type": "Neoplastic Process", "UMLS CUI": "C1641587", "Concept Status": "Inactive", "MODULEID": "900000000000207008", "Code in Source": "419602005"}, "retired": true, "source": "SNOMED", "names": [{"locale": "en", "locale_preferred": true, "name": "Carcinoma in situ of lacrimal drainage structure (disorder)", "name_type": "Fully Specified"}, {"locale": "en", "locale_preferred": false, "name": "Carcinoma in situ of lacrimal drainage structure (disorder)", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Carcinoma in situ of lacrimal duct", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Carcinoma in situ of lacrimal drainage structure", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Ca in situ of lacrimal drainage structure", "name_type": "Synonym"}]}, {"type": "Concept", "id": "115119004", "concept_class": "none", "datatype": "string", "extras": {"NHS Clinical Terms Version 3": "XU3iX", "UMLS Semantic Type": "Archaeon", "UMLS CUI": "C0524917", "Concept Status": "Active", "MODULEID": "900000000000207008", "Code in Source": "115119004"}, "retired": false, "source": "SNOMED", "names": [{"locale": "en", "locale_preferred": true, "name": "Family Thermococcaceae (organism)", "name_type": "Fully Specified"}, {"locale": "en", "locale_preferred": false, "name": "Family thermococcaceae", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Family thermococcaceae (organism)", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Thermococcaceae", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Family Thermococcaceae", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Family Thermococcaceae (organism)", "name_type": "Synonym"}], "parent_concept_urls": ["/orgs/International-Health-Terminology-Standard-Development-Organisation-IHTSDO-/sources/SNOMED/concepts/115212001/"]}, {"type": "Concept", "id": "365951004", "concept_class": "none", "datatype": "string", "extras": {"NHS Clinical Terms Version 3": "XUTVe", "UMLS CUI": "C1287494", "UMLS Semantic Type": "Finding", "Concept Status": "Active", "MODULEID": "900000000000207008", "Code in Source": "365951004"}, "retired": false, "source": "SNOMED", "names": [{"locale": "en", "locale_preferred": true, "name": "Finding of number of current sexual partners (finding)", "name_type": "Fully Specified"}, {"locale": "en", "locale_preferred": false, "name": "Number of current sexual partners - finding", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Finding of number of current sexual partners (finding)", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Finding of number of current sexual partners", "name_type": "Synonym"}], "parent_concept_urls": ["/orgs/International-Health-Terminology-Standard-Development-Organisation-IHTSDO-/sources/SNOMED/concepts/365950003/"]}, {"type": "Concept", "id": "365055005", "concept_class": "none", "datatype": "string", "extras": {"NHS Clinical Terms Version 3": "XUTHD", "UMLS Semantic Type": "Finding", "UMLS CUI": "C1286726", "Concept Status": "Active", "MODULEID": "900000000000207008", "Code in Source": "365055005"}, "retired": false, "source": "SNOMED", "names": [{"locale": "en", "locale_preferred": true, "name": "Finding related to ability to infer meaning (finding)", "name_type": "Fully Specified"}, {"locale": "en", "locale_preferred": false, "name": "Finding related to ability to infer meaning (finding)", "name_type": "Synonym"}, {"locale": "en", "locale_preferred": false, "name": "Finding related to ability to infer meaning", "name_type": "Synonym"}], "parent_concept_urls": ["/orgs/International-Health-Terminology-Standard-Development-Organisation-IHTSDO-/sources/SNOMED/concepts/365048008/"]} ], "extra_concepts": [], "Concepts compared": 466612, "missing_mappings": [], "Mappings compared": 126804, "extra_mappings": [] }

Note that there is extra output that lists the duplicate concepts, which is not included here but can be upon request.

[jamlung-ri]

LOINC has been validated using both the validator script and the hierarchy scan. No issues were found. SNOMED will still be reimported.