January
...
Differences found in package Comparison
...
1 new RefsetDescriptor records had their UUID's refreshed as expected
...
All x prefixes removed, and Production naming conventions added as expected
Plus, JSON file added in as planned:
> release_package_information.json
January 2022 Pre-Production Release (Version 5) vs January 2022 Pre-Production Release (Version 6) traceability
...
Differences found in package Comparison
...
RefsetDescriptor records refined to fix the AttributeDescription fields in line with the new MedDRA-SNOMED Map refset format spec (5.2.3.2 Simple Map to SNOMED CT Reference Set), as expected as fixed accordingly to https://ihtsdo.freshdesk.com/a/tickets/39074 :
- < 8ce7eb8e-2b0a-4075-b5f1-cffd8b325eef 20220131 1 816211006 900000000000456007 1193497006 900000000000505001 900000000000465000 1
- > 1cfcdb3e-8d8f-4b55-b953-4284b38daa19 20220131 1 816211006 900000000000456007 1193497006 900000000000500006 900000000000465000 1
- < e4e0dab9-a995-443a-b856-02e0426e222d 20220131 1 816211006 900000000000456007 1193497006 900000000000500006 900000000000461009 0
- > b4224fdc-c1ea-4c7b-8905-92fbf6ce9739 20220131 1 816211006 900000000000456007 1193497006 900000000000505001 900000000000461009 0
January 2022 Pre-Production Release (Version 4) vs January 2022 Pre-Production Release (Version 5) traceability
...
Differences found in package Comparison
...
New refsetDescriptor records added for the new MedDRA-SNOMED Map refset format, as expected:
- > 8ce7eb8e-2b0a-4075-b5f1-cffd8b325eef 20220131 1 816211006 900000000000456007 1193497006 900000000000505001 900000000000465000 1
- > e4e0dab9-a995-443a-b856-02e0426e222d 20220131 1 816211006 900000000000456007 1193497006 900000000000500006 900000000000461009 0
(NB We don't need to inactivate the old RefsetID refsetDescriptor records, as the community have always confirmed that the descriptor records should remain active even though the refset(s) in question are being inactivated)
**** ED THEN REPORTED AN ISSUE IN THE DESCRIPTOR RECORDS, SO THESE NEED UPDATING - https://ihtsdo.freshdesk.com/helpdesk/tickets/39074
...
ALL 6464 historical (now inactivated) records updated to confirm with the new format, as expected
The original format was incorrect (eg)
- id effectiveTime active moduleId refsetId referencedComponentId mapSource
- 89f1dec3-96a9-500b-9231-944250255860 20210131 1 816211006 816209002 10047801 190818004
- 89f1dec3-96a9-500b-9231-944250255860 20220131 0 816211006 816209002 190818004 10047801
- 7fe62b65-fa5a-5e78-a16a-1562784df08a 20220131 1 816211006 1193497006 190818004 10047801
But now it's corrected to show (eg)
- id effectiveTime active moduleId refsetId referencedComponentId mapSource
- 89f1dec3-96a9-500b-9231-944250255860 20210131 1 816211006 816209002 190818004 10047801
89f1dec3-96a9-500b-9231-944250255860 20220131 0 816211006 816209002 190818004 10047801 - 7fe62b65-fa5a-5e78-a16a-1562784df08a 20220131 1 816211006 1193497006 190818004 10047801
As requested by the community
...
2021 Production (PUBLISHED) vs January 2022 Pre-Production Release (Version 4) traceability
Differences found in package Comparison | Number of records impacted | Rationale |
---|---|---|
RefsetDescriptor files | 0 | NEEDS UPDATING WITH THE NEW REFSET ID FOR MEDDRA TO SNOMED (NB We don't need to inactivate the old RefsetID refsetDescriptor records, as the community have always confirmed that the descriptor records should remain active even though the refset(s) in question are being inactivated) |
Delta files (ALL) | All | ALL 16 Delta files removed to match International packaging standards, as expected |
MRCM files (x4) | All | 4x MRCM files added in to match International packaging standards, as expected |
Relationship Concrete Values files | All | Relationship Concrete Values files added in to match International packaging standards, as expected |
ModuleDependency files | 2 | Both records updated to 2022 dates, as expected |
Readme file | n/a | EffectiveTimes etc updated for Jan 2022, plus all Delta files removed, plus all MRCM + Relationship Concrete Values files added as expected |
SNOMED to MedDRA Map Snapshot files | 330 | 69 records inactivated 261 records added In TOTAL in the 2022 Snapshot record there are 3988 records:
This is as expected, as there were 3727 active records (plus 0 inactive) in the April 2021 release, less the 69 inactivated in April 2022 = 3658 active records expected in 2022. This matches the Delta file from WCI exactly (BUT NOT the stats in the WCI files - emailed WCI to see if they can get the stats working....). |
SNOMED to MedDRA Map Full files | 330 | 69 records inactivated 261 records added In TOTAL in the 2022 Snapshot record there are 4057 records:
This is as expected, as there were 3727 active records (plus 0 inactive) in the April 2021 release, plus the 261 activated + 69 inactivated in April 2022 = 4057 records expected in 2022. This matches the Delta file from WCI exactly (BUT NOT the stats in the WCI files - emailed WCI to see if they can get the stats working....). |
MedDRA to SNOMED Map Snapshot files | 13184 | 6464 records inactivated (12928 record changes in total due to new refsetID) 6719 added in for 2022 (but 6464 of these are the old records with the new refsetID, in the new format)
VERIFIED that using the same effectiveTime for the inactivation of the previous 6464 records is fine, as although we've got 2x records with the same effectiveTime + map pair in the snapshot files, this isn't a problem as they have new UUID's (eg)
This matches the stats in the WCI files exactly (6719 new records in 2022 Prod release), but this is to be expected as technically all those records are "new" for 2022 due to new format/refsetID. This matches the Delta file from WCI exactly:
This is as expected, as there were 6464 active records (plus 0 inactive) in the April 2021 release Snapshot, plus the 255 activated + 0 inactivated in April 2022 = 6719 records expected in 2022. |
MedDRA to SNOMED Map Full files | 13184 | 6464 records inactivated (12928 record changes in total due to new refsetID) 6719 added in for 2022 (but 6464 of these are the old records with the new refsetID, in the new format)
VERIFIED that using the same effectiveTime for the inactivation of the previous 6464 records is fine, as although we've got 2x records with the same effectiveTime + map pair in the snapshot files, this isn't a problem as they have new UUID's (eg)
What IS EXTREMELY CONFUSING HOWEVER, IS THAT THE OLD ORIGINAL ACTIVE RECORDS IN THE FULL FILE ARE IN THE WRONG FORMAT!!! (ie) the MapSource is in the referencedComponentId field, and the SNOMEDID is in the mapSource column!!! (eg)
THIS NEEDS FIXING SOMEHOW!!! Confirmed with Donna + Jane (on call on 22/3/22) that this should be fixed if possible, BUT if not possible in time for April then we can publish as it is, with a note in the Release Notes stating that a) the history will look strange, and so b) anyone needing to use the history should come and talk to us about how to do so... **** FIXED BY HACKING PREVIOUSLY PUBLISHED FILE TO PUT THESE 6464 RECORDS INTO THE NEW FORMAT, AND RUNNING THE SRS AGAINST THIS REFINED PACKAGE! This matches the stats in the WCI files exactly (6719 new records in 2022 Prod release), but this is to be expected as technically all those records are "new" for 2022 due to new format/refsetID. This matches the Delta file from WCI exactly:
This is as expected, as there were 6464 active records (plus 0 inactive) in the April 2021 release Full file, plus 6464 inactivation records for those original records, plus the 6719 new format records activated + 0 inactivated in April 2022 = 19647 records expected in 2022. |