OHDSI Home | Forums | Wiki | Github

How to represent family history

Vocabulary Users
#41

Thanks for noticing! Will take look.

It came from Themis group and it’s decision was as I posted above.

It seems like the stumbling block here is the ability of Usagi to recognize those concepts.
I don’t think that it’s necessary to expand “Maps to value” until they can be actually useful.

#42

So, is it the final decision?

This way we need to remap all such ICDs to support this decision

#43

Yep, correct. It’s 4210989 Family history with explicit context’ + value. We are almost compliant with it as we already have this structure in ICDs, just need to polish :slight_smile:

1 Like
#44

Hello,

My name is Wilma. I am an Intern in Big Data Vanderbilt University Medical Centre. I need to working on family history - for designated ICD codes, and I am new for working on this. Am I in the correct room to asking how the concept for the family history been established? I understand from the previous conversation that there are some concepts, but how could you find the code for the family history?

I appreciate everyone response for this. Thanks so much!

Best,
Wilma

#45

Welcome @wilma !

There are a couple different ways family history data can be coded in the CDM.

The first is with observation.observation_concept_id = 4210989, family history with explicit context, and the disease/condition is found in the observation.value_as_concept_id.

Or the data might be pre-coordinated as seen with this concept_id.

#46

Hi Melanie @MPhilofsky ,

Thanks so much for your response! I designted it as a concept_ID.

Thanks,
Wilma

1 Like
#47

Be careful with the Family history with explicit context. One of the descendant branch is No family history of. So when you pick the top-level guy you don’t actually say whether it’s happened or not :smile:

So it’s better to pick more specific, e.g. Family history of clinical finding or Family history with explicit context pertaining to mother.

2 Likes
#48

We need to create a convention around these.

1 Like
#49

Thanks @Alexdavv!

Were anyone of you understand if there is a way to indicate age of family member at diagnosis?
I am still confused working on mapping this family history to OMOP - still need to understand better the concept! :relieved:

Please if anyone of you already working on this longer so I may need your help as you have better understanding!

Thanks,
Wilma :grin:

#50

Even though some FH concepts have an indication that age is known (e.g.
FH angina female first degree age known), there’s no such indication that age can be recorded as the value.
Also, there’s no such convention in OMOP, so nobody will be looking the age in the value_as_number fields.

The only option that comes to my mind:
If you already have the relatives in the PERSON table (or you can create the records for them), you can calculate the dates and add the Condition records for them. Then just link persons to their relatives using the fact_relationship table.

1 Like
#51

Thanks so much @Alexdavv ! for your insight!

Best,
Wilma

#52

So let’s do the little surgery in SNOMED cutting this branch of

image
image1029×276 17.9 KB

then we can use “Family history with explicit context” as a generic concept as it covers relatives histories and we are not risking including negation concepts.

#53

The bigger problem is “the explicit context” can be different and it’s also a negative context. SNOMED uses this definition here and there.

And there’s a lot of confusing branches beneath we need to tackle:

I’d rather think about extensive clean-up and reorganization using OMOP Extension.
Also we need to decide where we allow pre-coordination and where we split up.

#54

you mean to create our own set of concepts? And not using the SNOMED at all, right?

#55

It’s rather deStandardize useless SNOMEDs and organizes the remaining ones into the new OMOP Extension hierarchy.

#56

I don’t like a mix here.
We can declare the OMOP Extension as a vocabulary to represent historyh flavors in OMOP and de-standardize the whole branch of SNOMED terms, mapping SNOMED concepts to OMOP Ext if they are equivalent.

#57

Vocab WG session on this?

#58

How can you establish a temporal relationship between two or more individuals for a given period of time? Where can I state a period between two dates? I need to relate COVID-19 infected individuals, who were in close contact with each other.

#59

@edomax:

OMOP is patient centric. One patient centric, that is. Relationships between patients is possible through FACT_RELATIONSHIP, but these are facts, not periods.

What is it you are trying to study? With what data?

#60

As a first step, I need to relate COVID-19 infected patients with their respective close contacts.

This relationship is for a certain period of time, since later the patients could have been infected again and have had a different group of close contacts to the first episode.

What we want to study is the performance of the traceability of COVID-19 patients and their close contacts.

t