search

OHDSI / CommonDataModel

Definition and DDLs for the OMOP Common Data Model (CDM)
https://ohdsi.github.io/CommonDataModel
896 stars 453 forks source link

How to store family history in the OBSERVATION table #656

Open jiawei-qian opened 7 months ago

jiawei-qian commented 7 months ago

How to store family history in the OBSERVATION table

CDM or THEMIS convention?

CDM

Table or Field level?

Table

Is this a general convention?

No

Summary of issues

  1. If these is an established best practice for representing family history in the OBSERVATION table? There seems 2 options for this family history. Image

Summary of answer

  1. Melanie https://forums.ohdsi.org/t/how-to-represent-family-history/3386/45 There are 2 ways: The first is the OBSERVATION.observation_concept_id = 4210989 (Family history with explicit context) and the disease/condition is found in the OBSERVATION.value_as_concept_id = 317009 (Asthma). Or the data might be the pre-coordinated with 4051241 (Family history of asthma).

  2. Erica Voss https://forums.ohdsi.org/t/how-to-represent-family-history/3386/9 RECOMMENDATION: If we have a matching concept in SNOMED, then use the respective observation_ concept_id. If we do NOT have a matching SNOMED code, then observation_concept_id should be ‘4210989 Family history with explicit context’ and value_as_concept_id should be the related procedure, condition etc.

    ACTION: Add as a convention under the OBSERVATION page.

  3. Alexander Davydov https://forums.ohdsi.org/t/how-to-represent-family-history/3386/47 Be careful with the 4210989 (Family history with explicit context). One of the descendant branch is 4051104 (No family history of). So when you pick the top-level guy you don't actually say whether it's happened or not. So it's better to pick more specific, e.g. 4167217 (Family history of clinical finding) or 4051255 (Family history with explicit context pertaining to mother).

Related links

OHDSI Forum: https://forums.ohdsi.org/t/how-to-represent-family-history/3386

Other comments/notes

From my side, I prefer to

  1. Firstly, use specific observation domain standard concept which represent the family history directly, e.g. 4051241 (Family history of asthma).
  2. Secondly, if I cannot find a specific standard concept directly, then use 4167217 (Family history of clinical finding) as OBSERVATION.observation_concept_id and 317009 as OBSERVATION.value_as_concept_id.

We have ETL Conversion projects in our organization. I used this logic for our OMOP CDM database. Yang, Qi is my colleague. I also discussed this logic with him before.

@MelaniePhilofsky, @clairblacketer, I am not sure if there is a convention on Family history data now. According to the discussion in this OHDSI Forum post, Melanie suggested the logic for how to store family history into OBSERVATION table. But is there any order for which way should be first? And if it is a good choice to replace 4210989 (Family history with explicit context) with 4167217 (Family history of clinical finding)?

MelaniePhilofsky commented 7 months ago

Moving this to the "Needs more work" column. The Vocabulary team has done more work on this topic, we need to collaborate with them before providing guidance.

MelaniePhilofsky commented 2 months ago

This topic has not been solved and is still under discussion. Please contribute your use cases and ideas to this forum post: https://forums.ohdsi.org/t/how-to-represent-family-history/3386