Death events in CDM V6

Ajit_Londhe · June 24, 2020, 3:19pm

Hi all,

In the CDM V6 wiki, the following convention is established for death:

The DEATH_DATETIME in the PERSON table should not be used as the way to find all deaths

select * from PERSON where death_datetime is not null should not be the practice

Rather, deaths should be found through the OBSERVATION table and the PERSON table is only used to determine which death date should be used in analysis

If the person table has the best available death dates for all dead patients, what is the consequence of simply querying from PERSON.death_datetime to find all dead patients? Certainly that query alone can’t get you death cause(s) or other context, but would it really be a faulty query if all you’re trying to do is identify who is dead and when?

Thanks,
Ajit

Christian_Reich · June 29, 2020, 11:23am

@Ajit_Londhe:

You are absolutely right. We are in the process of revising these definitions. Please hold on.

Whether or not your query is legitimate depends on whether death is captured with sufficient sensitivity. Has nothing to do with death per so.

Ajit_Londhe · June 29, 2020, 5:48pm

Thanks @Christian_Reich! I’ve opened a Git issue:

github.com/OHDSI/CommonDataModel

Rewrite death convention in V6

opened 02:46PM - 29 Jun 20 UTC

closed 02:39AM - 19 Feb 21 UTC

alondhe

Convention DEATH Documentation

Based on this forum thread, we may need to modify the conventions around queryin…g for death events in V6: https://forums.ohdsi.org/t/death-events-in-cdm-v6/11269 **Proposal:** The DEATH_DATETIME in the PERSON table can be used to identify patients who died and when they most likely died. However, to obtain full context around death events, use the OBSERVATION table and join to the CONDITION_OCCURRENCE table to find death cause(s).