Investigate submission of investigational drugs which do not exist in RxNorm

[hknahal]

Example: "Dovitinib" and "Celastrol" do not exist in RxNorm.

Investigate:

• Whether drug exists in other drug databases. For example "Dovitinib" exists in other databases (https://go.drugbank.com/drugs/DB05928#external-links).
• If drug exists in other database, do they have a publicly accessible API to query drug name for accession identifier?

[hknahal]

Proposed changes that were implemented in MOH data model as well.	Field name	Data Tier	Attribute	Type	Description	Permissible Values
drug_reference_database	Core	Required	Indicate the reference drug database where drug name is found.	Text	RxNorm, PubChem, NCI Thesaurus
drug_name	Core	Required	Indicate the drug name as it exists in the database specified in the drug_reference_database.	Text
drug_reference_identifier	Core	Required	Indicate the identifier from the "drug_reference_database" for the drug.	Text

Having generalized fields that are not specific to just RxNorm enable addition of new drug reference databases in the future, should they be required.

Similar to RxNorm, we can use an API for PubChem and NCI Thesaurus to validate the submitted "drug_name" and "drug_reference_identifier" fields.

To validate drug_name and drug_reference_identifier from PubChem:

• We can use PubChem's API

Example:	Field name	Value	Validation
drug_reference_database	PubChem
drug_name	Celastrol	https://pubchem.ncbi.nlm.nih.gov/rest/pug/compound/name/Celastrol/cids
drug_reference_identifier	122724	https://pubchem.ncbi.nlm.nih.gov/compound/122724

• Pass "drug_name" value "Celastrol" to API: https://pubchem.ncbi.nlm.nih.gov/rest/pug/compound/name/Celastrol/cids. This will return a list of CIDs that match this drug name, in this case 122724:

  <IdentifierList xmlns="http://pubchem.ncbi.nlm.nih.gov/pug_rest" xmlns:xs="http://www.w3.org/2001/XMLSchema-instance" xs:schemaLocation="http://pubchem.ncbi.nlm.nih.gov/pug_rest https://pubchem.ncbi.nlm.nih.gov/pug_rest/pug_rest.xsd">
  <CID>122724</CID>
  </IdentifierList>

• Verify if CID(s) returned from API match "drug_reference_identifier". In this case, it does match.

To validate drug_name and drug_reference_identifier from NCI Thesaurus:

• Download latest tsv dump file of NCI Thesaurus from https://evs.nci.nih.gov/evs-download/thesaurus-downloads. Can load this file in DB and search on drug_reference_identifier and drug_name.

Example:	Field name	Value
drug_reference_database	NCI Thesaurus
drug_name	Allogeneic iC9/CAR19/IL15-transduced CB-NK Cells
drug_reference_identifier	C148215

Search "code" column for drug_reference_identifier" and check if "drug_name" matches any of the terms that exist in the "synonym" column.

[edsu7]

Potential subtickets

Potentially work for DR11

• data migration
• developer : update validation scripts
  • this includes parsing new databases/APIs
• new dictionary field?
  • extend databases based on MONSTAR's drug list
  • add new field to capture nonnamed drugs (i.e. honey)
• Bioinfo to confer with MONSTAR for # of records affected