Age is a common, basic piece of information that poses several challenges.
Most of the time, ‘Age' stands for ‘Current Chronological Age of the Subject’, as computed from
a Date of Birth and the “current” timestamp. However, ‘Age’ can be computed relatively to any other relevant event in the subject’s history (past, present or future).
It is a contextual, derivative data element which involves non-trivial mathematics and poses challenges in terms of granularity of representation. Its proper modeling varies depending on the use cases
Aspects to be considered:
Clarifying the semantics of the concept - Current Chronological Age, vs Estimated Age, vs Estimated Age via Bone Age, etc…
Is the concept “code” captured in the schema vs in the data
Granularity of the (complex) data type that holds an age:
One unit, vs break it down in components (days, weeks, months, years, decades)
Rounding of the value
Unit conversions: implicit or explicit?
Utility of birthdate and other ‘anchor’ events
calculating age at event as date minus birthdate
calculating date of event as age of event plus birthdate
The anchor event (e.g. birth) may be known only to some granularity
Provenance/Pedigree of ‘Age’
computed
asserted (e.g. via questionnaire)
stated (e.g. by patient)
predicted/estimated
Based on the aspects above, different modeling patterns may be fit for purpose
Examples of approaches to modeling “““Age”””
Background / Related References:
Age in FHIR
Age in CQL
….