Guidance on how to map multiple alternative causes of death

[judithb (Judith)]

Hello,
We are preparing our source data for mapping to the OMOP CDM and are unsure how to handle death records when the source data contains more granular information (e.g., multiple alternative causes of death) than what can be stored in the death table. Should these causes instead be represented in the condition_occurrence table, flagged as main or alternative causes of death using the condition_status_concept_id?" Thank you!

[Christian_Reich]

Hi @judithb:

Welcome to the family.

Two answers:

1. Yes, causes for death are conditions, and there is nothing wrong with storing them there. Yes, put Death Certificate into condition_type_concept_id.
2. You can have more than one record in the DEATH table, as long as THE DATE is THE SAME.

CR

[MPhilofsky (Melanie Philofsky)]

Wait! @Christian_Reich OHDSI, specifically Themis and the CDM WG, have not approved the solutions you lay out:

@Christian_Reich

Two answers:

1. Yes, causes for death are conditions, and there is nothing wrong with storing them there. Yes, put Death Certificate into condition_type_concept_id.
2. You can have more than one record in the DEATH table, as long as THE DATE is THE SAME.

This continues to be an open question per CDM GitHub issue.

I know you know, so please take this quick lecture light-heartedly

We can’t enable reliable, reproducible research if we everyone makes up their own rules on how and where to insert data to the OMOP CDM. Put it here or there is not good advice or an appropriate solution. We need to declare one rule with conventions for all to follow.

@judithb Please comment on the CDM GitHub issue linked above. Include your use case and the importance of having the issue resolved. It has lingered way too long.