Cset management: Joe 2024/03

[joeflack4]

Overview

Joe to includeDescendants on new versions of Siggie's csets from this spreadsheet, and report results.

Results

I'm not too sure what to look for and tell you guys here, but:

1. Hypothesis confirmed: Good news. So far results are consistent with our hypothesis. The "...is newly added" entries in "active pending changes" are simply the new concepts that would be added if all of includeDescendants were set to true.
2. valid_start_date: FYI I checked the "valid start date"s of the new concepts. For case (1), they have both been valid as of May 21, 2023, and for case (2), Feb 16, 2022. All other "valid start date"s for all other concepts in these csets are much older, with the newest one being 2016 I believe.
3. TermHub usage: I did 2 comparisons; for each cset, comparing the base version (v1) with the new versions I created. UI (content and formatting) was as I expected, with the exception of the "Concepts in set but not linked to others" in Procalcitonin, ng/mL. See below.
   Concepts in set but not linked to others

I don't know if I'm reading this incorrectly, or if it's a bug. I suspect the latter. Shouldn't the two concepts shown here be roots?

Also, what is this feature trying to show? Shouldn't every case in comparing a concept/expression between multiple csets be covered by 1 of the 4 cases already handled by the UI?

• Concept is in the definition but not the expansion
• Concept is in the expansion but not the definition
• Concept is in the definition and the expansion
• Concept is not in this concept set at all

1. NT pro BNP, pg/mL (561166072)

• Expression items: 14
• n concepts in v1 (all includeDescendants=false): 14
• n concepts in new version (all includeDescendants=true): 16
• n expected new concepts (Active pending changes entries: "...is newly added": 2
• n actual new concepts: 2

2 new concepts added, they are CPT (other concepts in the cset were either LOINC or SNOMED only), and they are much more specific, strange label like w/ whole path in it, and they have 0 records:

• Cardiology (cardiovascular disease), analysis of 4 proteins (NT-proBNP, osteopontin, tissue inhibitor of metalloproteinase-1 [TIMP-1], and kidney injury molecule-1 [KIM-1]), plasma, algorithm reported as a risk score for major adverse cardiac
• Pediatrics (vasculitis, Kawasaki disease [KD]), analysis of 3 biomarkers (NT-proBNP, C-reactive protein, and T-uptake), plasma, algorithm reported as a risk score for KD

https://icy-ground-0416a040f.2.azurestaticapps.net/cset-comparison?codeset_ids=561166072&codeset_ids=478183721

2. Procalcitonin, ng/mL (610397248)

• Expression items: 6
• n concepts in v1 (all includeDescendants=false): 6
• n concepts in new version (all includeDescendants=true): 7
• n expected new concepts (Active pending changes entries: "...is newly added": 1
• n actual new concepts: 1

1 new concept:

• Procalcitonin [Moles/volume] in Serum or Plasma

Doesn't have any records.

https://icy-ground-0416a040f.2.azurestaticapps.net/cset-comparison?codeset_ids=610397248&codeset_ids=432635786

---

Probably don't need to contact author because 0 records; doesn't affect results.

[joeflack4]

@Leskena2 @Sigfried Done. Let's examine results at scrum.

[joeflack4]

concepts in set but not in others = orphans