IDMP — Substances, Products & Packaging

The five ISO standards, the identifier family (MPID/PhPID/PCID), and the faceted product model.

0/4 done

Theory

Five standards, one coherent identity system

IDMP isn't one document — it's a family of five ISO standards that interlock. Knowing which does what is the day-job literacy a regulatory-data role expects:

  • ISO 11238 — Substances. Identifies the substance itself (the active or inactive material) with a Substance ID.
  • ISO 11239 — Pharmaceutical dose forms, units of presentation, routes of administration. The controlled vocabularies for how a product is presented and given.
  • ISO 11240 — Units of measurement. Standardised units (anchored to UCUM) so strengths and quantities compare across regions.
  • ISO 11615 — Medicinal products. The authorised product as a regulator licenses it, identified by an MPID.
  • ISO 11616 — Pharmaceutical products. The scientific product (substances + strength + dose form, independent of brand or packaging), identified by a PhPID.

Licensing note: the ISO IDMP texts are paywalled. Everything below is an illustrative model of the structure, with example identifiers — not reproduced standard content.

The identifier family — why so many?

Each identifier answers a different question, and conflating them is exactly the safety bug IDMP exists to prevent:

  • MPIDthis authorised product, in this market (brand, marketing-authorisation holder, country).
  • PhPIDthe science: same substances + strength + dose form, regardless of brand. Two brands of the same generic share a PhPID but have different MPIDs.
  • PCIDthis packaged configuration (the 28-tablet blister vs the 100-tablet bottle).
  • Substance IDthe material a product contains.

The faceted model

A medicinal product hasPharmaceuticalProduct a scientific product, which containsSubstance one or more substances at a strength (a quantity in ISO 11240 units), and is packaged into configurations each with its own PCID. Reasoning over this graph is what lets a regulator answer 'every authorised product, in any country, that contains substance X above strength Y.'

Use Case Example: A safety signal implicates a specific active substance. With IDMP, an agency walks Substance ID → every PhPID that contains it → every MPID (brand/market) carrying that PhPID → every PCID on shelves — turning a vague 'recall things with drug X' into an exact, packaging-level worklist.

Analogy

IDMP's identifiers are the ISBN, the edition, and the physical copy of a book. The work ('Hamlet') is the science — the PhPID. A specific published edition by a specific publisher in a specific country is the authorised product — the MPID. The boxed set vs single paperback is the packaging — the PCID. Libraries learned long ago that you need all three levels; pharmacovigilance needs them for the same reason — to recall exactly the right thing.

Substance → PhP → MP → package

Click a node to focus its neighbourhood · drag to pan · scroll to zoom

The faceted IDMP graph

Substance feeds the pharmaceutical (scientific) product, which a medicinal (authorised) product realises and packaging configures — each facet with its own identifier.

Reflect

IDMP is faceted identity taken to its logical end: the substance, the science, the authorised product and the box on the shelf are four questions, so they get four identifiers. The discipline is annoying right up until the day you must recall exactly one of them worldwide.

  • Which 'one entity' in your domain is really several facets that deserve separate identifiers?
  • How would a shared scientific identifier (PhPID-style) change how you deduplicate brands/variants today?

Reading in progress · 0 of 4 activities done