TODO - Manuscript Task: Verify Peanut Allergy Phenotype

[callahantiff]

Creating as a separate issue to break-up issue #2

Finalizing the Peanut Allergy Phenotype

Case Criteria:

• CASE TYPE 1
  • Presence of ANY mention of "peanut" in any section of EMR (keyword)
  • Presence of quantitative immunoassay testing performed (RAST) (lab string)
    OR
  • Presence of Allergen specific IgE, quantitative (Procedure code)

• CASE TYPE 2
  • Presence of ANY mention of "peanut" in any section of EMR (keyword)
  • Presence of >1 of various allergy testing procedures (CPT code)
  • Percutaneous (scratch, prick, or puncture test) (CPT code)
  • Intracutaneous test (intradermal test) (CPT code)
  • Food ingestion challenge (CPT code)

Control Criteria:

• No control group defined

---

Questions:

• No control group defined, should we try to create one?

• Is Peanut RAST (SNOMED: 388500008) the best term to meet the criteria for "Presence of quantitative immunoassay testing performed (RAST)"?

---

@tdbennett, can you please advise on these questions?

[tdbennett]

1. I'm ethically against control groups

2. This is probably a dumb question, but why not LOINC 81991-2?

[callahantiff]

Thanks Tell -

• Great, I updated the documentation.
• We don't have that specific test, but we do have these:

Row	LOINC
1	58779-0: Peanut recombinant (rAra h) 1 IgE Ab [Units/volume] in Serum
2	64965-7: Peanut recombinant (rAra h) 9 IgE Ab [Units/volume] in Serum
3	58778-2: Peanut recombinant (rAra h) 2 IgE Ab [Units/volume] in Serum
4	58777-4: Peanut recombinant (rAra h) 3 IgE Ab [Units/volume] in Serum

Do you want to use these instead of SNOMED_388500008: Peanut RAST?

[tdbennett]

I honest have no idea whether the LOINC or the SNOMED gets used more for these tests. Why not the union?

[callahantiff]

Sounds good to me, I actually think that we might want to query for any measurement that contains "%peanut rast %" or "%peanut recombinant %". This gives us:

Row	source_code	source_name	source_domain_id
1	15917-8	Peanut IgE Ab RAST class [Presence] in Serum	Measurement
2	21447-8	Peanut IgG Ab RAST class [Presence] in Serum	Measurement
3	81990-4	Peanut recombinant (rAra h) 1 IgE Ab RAST class [Presence] in Serum	Measurement
4	81991-2	Peanut recombinant (rAra h) 2 IgE Ab RAST class [Presence] in Serum	Measurement
5	81992-0	Peanut recombinant (rAra h) 3 IgE Ab RAST class [Presence] in Serum	Measurement
6	87719-1	Peanut recombinant (rAra h) 5 IgE Ab RAST class [Presence] in Serum	Measurement
7	87718-3	Peanut recombinant (rAra h) 6 IgE Ab RAST class [Presence] in Serum	Measurement
8	81982-1	Peanut recombinant (rAra h) 8 IgE Ab RAST class [Presence] in Serum	Measurement
9	81994-6	Peanut recombinant (rAra h) 9 IgE Ab RAST class [Presence] in Serum	Measurement

• @tdbennett are you OK with us including these labs?

[tdbennett]

Great

…

On Apr 1, 2019, at 2:45 PM, Tiffany J. Callahan ***@***.***> wrote: Sounds good to me, I actually think that we might want to query for any measurement that contains "%peanut rast %" or "%peanut recombinant %". This gives us: Row source_code source_name source_domain_id 1 15917-8 Peanut IgE Ab RAST class [Presence] in Serum Measurement 2 21447-8 Peanut IgG Ab RAST class [Presence] in Serum Measurement 3 81990-4 Peanut recombinant (rAra h) 1 IgE Ab RAST class [Presence] in Serum Measurement 4 81991-2 Peanut recombinant (rAra h) 2 IgE Ab RAST class [Presence] in Serum Measurement 5 81992-0 Peanut recombinant (rAra h) 3 IgE Ab RAST class [Presence] in Serum Measurement 6 87719-1 Peanut recombinant (rAra h) 5 IgE Ab RAST class [Presence] in Serum Measurement 7 87718-3 Peanut recombinant (rAra h) 6 IgE Ab RAST class [Presence] in Serum Measurement 8 81982-1 Peanut recombinant (rAra h) 8 IgE Ab RAST class [Presence] in Serum Measurement 9 81994-6 Peanut recombinant (rAra h) 9 IgE Ab RAST class [Presence] in Serum Measurement @tdbennett are you OK with us including these labs? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[callahantiff]

We have addressed all open questions; closing this issue for now.