This issue will be updated continuously as we realise things that needs to be explained/addressed/justified in the documentation of the data. Things won't be removed but ticked off when included in the documentation whose contents live in the docs branch.
[ ] MEASUREMENT.value_as_concept_id will use just 45878416 Blood, and users who desire to distinguish venous from arterial blood, will need to use the SPECIMEN table.
[ ] type_concept_id for data coming from LABKA and BCC is 32856
[ ] person from source are also filtered so that the d_status_hen_start is higher than birthday and beginning of registry. This will avoid including patients who died before the registry starts and also patients who do not have enough follow-up
[ ] When using quant clinical drugs, we always use the one with the smallest volume and that is shared across concentrations for the same ingredient. If no such concept exists, we resort to the smallest volume on a per-concentration basis
[ ] These data must be analysed with Central European locale (e.g. Europe/Copenhagen), otherwise the date parts of start_date and start_datetime won't match; the same goes for end_date and end_datetime)
[ ] If a medication is registered in both 3- files and 8-/9-files, we assume these are different drugs even if they might actually different as there's no way to disentangle this in the ETL pipeline. These kinds of decision are deferred to the analytic phase.
[ ] Conditions only have eras if they have a valid era_lookback_interval value in concept_lookup_stem. This is because we want finer control, and we don't want to rely too much on environment variables. Drugs use a default era_lookback_interval, which can be controlled by an environment variable although it's unlikely to ever be.
This issue will be updated continuously as we realise things that needs to be explained/addressed/justified in the documentation of the data. Things won't be removed but ticked off when included in the documentation whose contents live in the docs branch.
MEASUREMENT.value_as_concept_idwill use just 45878416 Blood, and users who desire to distinguish venous from arterial blood, will need to use theSPECIMENtable.type_concept_idfor data coming from LABKA and BCC is 32856Europe/Copenhagen), otherwise the date parts ofstart_dateandstart_datetimewon't match; the same goes forend_dateandend_datetime)