Observational Health Data Sciences and Informatics

link to github issue

• Requester: Klaus Bonadt
• Discussion: here.

Proposal for conventions to achieve consistent entries in the attribute quantity of the DRUG_EXPOSURE table and entries in the DRUG_STRENGTH table. Eventually, we will be able to calculate the total amount of the active ingredient by

• DRUG_EXPOSURE.quantity x DRUG_STRENGTH.numerator_value [numerator_unit_concept_id] (for concentrations) or
• DRUG_EXPOSURE.quantity x DRUG_STRENGTH.amount_value [amount_unit_concept_id] (for pieces, e.g. tablets)

For drugs with a dose release over time (e.g. Patches) we will be able to calculate the hourly rate by

• DRUG_STRENGTH.numerator_value [numerator_unit_concept_id]

1. Roughly speaking, the value for quantity represents the quantity of the drug in the denominator unit of the corresponding entry (or entries) in the DRUG_STRENGTH table. More precisely,
   1. For drugs dispensed in “pieces” (e.g. tablets or puffs of an inhaler) the quantity refers to this number of pieces. (Thus, quantity has no unit and the denominator unit in the DRUG_STRENGTH table is either empty or {actuat}).
   2. For drugs with a concept_id denoting a quantified drug the quantity refers to a fraction/multiple of the whole pack. (The quantity has no unit, the corresponding entry/entries in the DRUG_STRENGTH table refer to the concentration with the total amount in the denominator).
   3. For all other drugs quantity refers to the total amount of the product measured in mL or g. (Thus, the unit of the quantity is either g or mL – both units can be used interchangeably.
2. For products with a dosage release rate (for example patches), the quantity refers to the number of pieces. (The denominator unit of the corresponding entry in the DRUG_STRENGTH table is hour).
3. Compounding: For each compound one entry will be added to the DRUG_EXPOSURE table. This enables us all the above discussed options to provide quantity and strength. Even a compounding of multi-ingredient products would be possible.

Probably, nothing has to be changed. But we should cross-check the following assumed properties of the DRUG_STRENGTH table coming from RxNorm and apply these for future enhancements:

• For drugs dispensed in “pieces”, all corresponding entries must have either an empty denominator unit or {actuat} in the denominator unit.
• For drugs with a dose release over time, we use hour in the denominator.
• For all other drugs, the denominator unit refers to mass or volumes, measured in mL or mg.

With the above mentioned conventions the fields effective_drug_dose and dose_unit_concept_id can be removed. They don’t support multiple ingredients anyway.

In order to support quantified drugs we need an additional field denominator_value. For compounding we would need to add records to DRUG_STRENGTH for all compounds (drug_concept_id=ingredient_concept_id) with the mg concept in the denominator_unit_concept_id for all "dry" compounds and mL for the "wet" ones. We could steal the knowledge what is what from RxNorm.

1. We could agree on using mg instead of g for the quantity in case 1c. However, this would require additional conventions and steps for the ETL process: We would use mL for liquids and products with different denominator units (for example 42799258 “Benzyl Alcohol 0.1 ML/ML / Pramoxine hydrochloride 0.01 MG/MG Topical Gel”). However, we would use mg for products which exclusively use a mass unit. Thus, we would need to check all corresponding entries in the DRUG_STRENGTH table beforehand in order to know whether we can apply mg or mL. I don't recommend this.
2. We could use an additional field quantity_unit_concept_id to denote the unit. However, this would also require additional sets of conventions to avoid incompatible entries between quantity_unit_concept_id and denominator_unit_concept_id. I don't recommend this.
3. In the DRUG_STRENGTH table we could allow additional units for masses (besides mg), or volumes (besides mL) or time (besides h). However, this would add not only additional complexity for the conversions but also for the checks to decide whether a DRUG_STRENGTH entry belongs to a dose release over time, to a mass or a volume unit. I don't recommend this.
4. We might want to distinguish between drugs used in compounding and normal drugs dispensed on the same day. In the DRUG_EXPOSURE table we could use a compounding_sequence (holding a sequence number for each compounding on the same day for the same patient) or a compound_sequence (holding a sequence number for each compound. The sequence will restart from 1 for all subsequent compounding). I do recommend this, preferably the latter one.