There is no 'concept set' for a Observation Period criteria:
This may be a symptom of trying to re-use criteria elements across all criteria types, but we should avoid this 'code reuse' temptation: In the CDM, ConditionOccurrence, DrugExposure, DrugEra, ProcedureOccurrence, etc etc a re all represented as separate tables so even tho they might seem like there is shared fields between them, they are distinct entities and we should not try to represent the object model or UI elements to imply that there is some sort of hierarchy or shared behavior between them. We should avoid patterns like 'ConditionOccurrence, drug exposure and Procedure Occurrence' share X fields so these can have some common attrbitue collcection, while DrugEra and ConditionEra share Y fields so they can have some common attribute collections. This would lead to a complex web of making simlilarities between objects where they conceptually don't exist, only coincidently, and the object model of one of these criteria types can change independently of another (such as introducing date/time of a proceedure while leaving condition diagnosis at the date level).
There is no 'concept set' for a Observation Period criteria:
This may be a symptom of trying to re-use criteria elements across all criteria types, but we should avoid this 'code reuse' temptation: In the CDM, ConditionOccurrence, DrugExposure, DrugEra, ProcedureOccurrence, etc etc a re all represented as separate tables so even tho they might seem like there is shared fields between them, they are distinct entities and we should not try to represent the object model or UI elements to imply that there is some sort of hierarchy or shared behavior between them. We should avoid patterns like 'ConditionOccurrence, drug exposure and Procedure Occurrence' share X fields so these can have some common attrbitue collcection, while DrugEra and ConditionEra share Y fields so they can have some common attribute collections. This would lead to a complex web of making simlilarities between objects where they conceptually don't exist, only coincidently, and the object model of one of these criteria types can change independently of another (such as introducing date/time of a proceedure while leaving condition diagnosis at the date level).