How to populate month_of_birth when it is not provided in the source

[clairblacketer]

How to populate month_of_birth when it is not provided in the source

CDM or THEMIS convention?

Themis

Is this a general convention?

No

Summary of issues

• Month of Birth is not required in the CDM, but it is a component of birth_datetime. The CDM conventions has guidance listed for how to infer month_of_birth in the birth_datetime field but not for the month_of_birth field itself.

Summary of answer

• If month_of_birth is null or if day_of_birth AND month_of_birth are both null and the person has records during their year of birth then use the date of the earliest record, otherwise use the 15th of June of that year.

Related links

• CDM Website: https://ohdsi.github.io/CommonDataModel/cdm54.html#person

Other comments/notes

• @clairblacketer @MelaniePhilofsky would appreciate your feedback on if/how we handle this one. I will be opening a similar one for day_of_birth.

[clairblacketer]

@burrowse The convention is correct for how we typically assign month at birth. I usually reserve this for kids and babies for whom we observe records during their year of birth, as you note. This is helpful for mother-child linkage algorithms, perhaps we put a note to that effect?

[burrowse]

@burrowse The convention is correct for how we typically assign month at birth. I usually reserve this for kids and babies for whom we observe records during their year of birth, as you note. This is helpful for mother-child linkage algorithms, perhaps we put a note to that effect?

Okay, great! That makes sense to me! This was mostly making sure that we were aligned across the month_of_birth, day_of_birth and birth_datetime field in terms of how to infer a value as a convention.

[clairblacketer]

Issue # and location

• Themis issue #94

Issue summary

How should month_of_birth be populated when it is not provided in the source data? Month of Birth is not required in the CDM, but it is a component of birth_datetime.

Convention type

Table

CDM table

person

CDM field

month_of_birth

Links to issue discussion

NA

Provenance of data

General

The ratified convention

Month of Birth is not required in the CDM, but it is a component of birth_datetime. For data sources that provide the precise date of birth, the day should be extracted and stored in this field. For data sources which lack this data point, but have a use case to infer the information (i.e. This is helpful for mother-child linkage algorithms), use the following heuristic:

• If month_of_birth is null or if day_of_birth AND month_of_birth are both null and the person has records during their year of birth then use the date of the earliest record, otherwise use the 15th of June of that year.

Date of ratification/published

NA

Downstream implications

No

Link to DQD check

NA

Related conventions/further information

NA

#Tags birth_date, month_of_birth

[MelaniePhilofsky]

Edited to match wording of day_of_birth issue.

[clairblacketer]

https://github.com/OHDSI/Themis/pull/150

[clairblacketer]

reopening for discussion as we may not need to give a convention for a not-required field

tagging @cgreich @MelaniePhilofsky

[MelaniePhilofsky]

Issue has been resolved.