Revisit adding VALUE_AS_DATE to the OBSERVATION table

[clairblacketer]

This was added in v6.0, do we want it in an upcoming version? Potentially related to "history of" conditions

[MaximMoinat]

I looked up the use case for the value_as_date(time) for a 'History of'. It was discussed on the forums, but we settled on using 'enriched' History Of concepts. So for that use case we don't need a value_as_date anymore.

And we need an analysis use case (yes, I know, I am just copying Christian...). e.g. when would we use value_as_date to construct a cohort definition?

[MaximMoinat]

Just came across this use case on the forums. To store the estimated (future) delivery date. https://forums.ohdsi.org/t/mapping-question-of-the-week-how-should-i-model-this-pregnancy-data/16062

Regarding the analysis use case, I can imagine one where we can determine early and late births. The again, this is likely also to be stored as a separate observation, e.g. 'Preterm birth'.

[MelaniePhilofsky]

The University of Colorado is participating in a non-OHDSI OMOP CDM network study which requires a date as a value in the Observation table.

Use case: Estimated Date of Deliver (EDD) is needed to determine if a pregnant person were diagnosed with high blood pressure before or after conception. High blood pressure in non-pregnant persons, known as hypertension, is much different disease process than high blood pressure which develops during pregnancy, known as preeclampsia. The EDD will be utilized in relation to the date of high blood pressure conditions and measurements to determine if the high blood pressure was observed before or after the beginning of pregnancy.

OBSERVATION.date is "The date of when the Observation was obtained". We also need OBSERVATION.value_as_date to identify the date as a value for the observation_concept_id. observation_concept_id = and the OBSERVATION.value_as_date = , which can be in the future.