search

EHDEN / ETL-UK-Biobank

ETL UK-Biobank
https://ehden.github.io/ETL-UK-Biobank/
12 stars 4 forks source link

Improve CTV3 mapping with NHS-provided SNOMED mappings #330

Closed MaximMoinat closed 2 years ago

MaximMoinat commented 2 years ago

We can use existing CTV3 mappings to SNOMED from the NHS.

There might be one to many mappings, then we take the simplest or 'assured' one.

https://forums.ohdsi.org/t/read-ctv3-vocabulary/7557/3

https://isd.digital.nhs.uk/trud/users/authenticated/filters/0/categories/1/items/98/releases https://isd.digital.nhs.uk/trud/users/authenticated/filters/0/categories/8/items/9/releases

MaximMoinat commented 2 years ago

One-to-many mappings. e.g. to both an 'Observational Entity' and a 'Procedure' class. Or both observation and measurement domain.

Example: XE25y,Urine paracetamol level

Maps to: 37392428,Urine paracetamol concentration,Measurement,Observable Entity 4211691,"Acetaminophen measurement, urine",Measurement, Procedure

@vpapez Do you have a suggestion which to choose?

vpapez commented 2 years ago

How many of them are there? Is this something we could review manually one-by-one or should we rather come up with some general approach?

MaximMoinat commented 2 years ago

There are 3500 of them. So a general rule would be preferred.

MaximMoinat commented 2 years ago

Every CTV3 code has one or more terms. We only map the Preferred (P) term, which can have a different meaning than the others. For example two of the six terms of code Xa9Bs (DVT).

ctv3_conceptid ctv3_termid ctv3_termtype snomed concept_id snomed_name
Xa9Bs Y20L6 P 443537 Deep venous thrombosis of lower extremity
Xa9Bs Y20Kh S 4133004 Deep venous thrombosis

The preferred term here is more specific than the others. In our mapping, all Xa9Bs codes will be mapped using the SNOMED of this more specific (preferred) term. This is adding a level of granularity that we don't have.

Ideally we also use the CTV3 term to do the mapping. @vpapez Do we have the CTV3 term in the source data?

vpapez commented 2 years ago

Let's go with the prefered one. It will be consistent with the rest of the mapping. Also, when I checked the DVT definition in AESI study, both these concepts are included.