About manufactured dose forms and clinical drugs

SNOMED Groups Drugs Extension User Group
1

Dear All,
Could you please share how you manage your national drug extension?
Do you apply specific guidelines or rules for the creation of clinical drugs and for their mapping to pharmaceutical products (brand names)? If so, would you be willing to share these procedures?
We have noticed some potential issues in the content of the international edition and would appreciate your insights.

  1. For injectable drugs, in some cases, the manufactured dose form includes specific details regarding the preparation methods prior to drug administration, whereas in other cases, such information is not provided:
  • The dosage form “Concentrate for conventional release solution for infusion” is available in SNOMED CT: the solution must not be administered to the patient before it has been properly diluted, and this requirement is clearly stated in the description of the manufactured dosage form (1237267003 |Concentrate for conventional release solution for infusion (dose form)|.
  • By contrast, the dosage form “Powder for concentrate for solution for infusion” is not included in SNOMED CT.
    The designation of the manufactured dose form, “Powder for concentrate for solution for infusion,” implies a two-step preparation process prior to administration:
    • First step: reconstitution
    • Second step: dilution of the reconstituted solution
      This distinction is critical from a safety standpoint, as the reconstituted solution must not be administered to the patient before it has been properly diluted.
  1. Some clinical drugs seems to be duplicates:
    For example : there are 3 clinical drugs with same strength for 430153008 |Product containing ganciclovir in parenteral dose form (medicinal product form)|
  • 1237153003 |Product containing precisely ganciclovir (as ganciclovir sodium) 500 milligram/1 vial powder for conventional release solution for infusion (clinical drug)|
  • 324781006 |Product containing precisely ganciclovir (as ganciclovir sodium) 500 milligram/1 vial powder for conventional release solution for injection (clinical drug)|
  • 1237154009 |Product containing precisely ganciclovir (as ganciclovir sodium) 500 milligram/1 vial powder for conventional release solution for infusion and/or injection (clinical drug)|
    But sources indicate that “Ganciclovir must be administered by intravenous infusion over 1 hour at a concentration not exceeding 10 mg/mL. Do not administer by rapid or bolus intravenous injection because the resulting excessive plasma levels may increase the toxicity of ganciclovir.”
    “Do not administer by intramuscular or subcutaneous injection because this may result in severe tissue irritation due to the high pH (~11) of ganciclovir solutions”
2

Hi @Elisase

I can add a bit of information here:

  1. In the Editorial Guide under Pharmaceutical Dose Form Pharmaceutical Dose Form | Specifications SNOMED CT Editorial Guide | SNOMED International Documents
    Out of Scope section: Concepts in pattern “x for y for z” (e.g. Powder for concentrate for dispersion for infusion)
  2. 324781006 |Product containing precisely ganciclovir (as ganciclovir sodium) 500 milligram/1 vial powder for conventional release solution for injection (clinical drug)| has a date of 2018. This is when the new drug model was implemented some of it by batch automation. So 2002 concept 376055000 |Ganciclovir sodium 500mg powder (product)| was inactivated and replaced by 324781006 |Product containing precisely ganciclovir (as ganciclovir sodium) 500 milligram/1 vial powder for conventional release solution for injection (clinical drug)|. If this concept is erroneous please submit a CRS request for inactivation. Many thanks Nicki
1 Like
3

Hi @Elisase

AMT takes a pragmatic approach to dose form modelling, where clinically real differences are not always represented if they are ambiguous, inconsistently available, or do not add value for common use cases. In practice, dose form is often generalised (e.g. “injection” or “powder for injection” instead of more specific variants), reflecting the manufactured dose form and the level of detail needed for product identification and grouping rather than full preparation or administration accuracy.

Although distinctions such as “powder for injection” versus “powder for solution for infusion” are clinically meaningful, this would be considered knowledge about the preparation of a dose form to administer a patient, and would be supported by clinical knowledge, dose base prescribing, and decision support systems. AMT previously had more granular dose forms, however after extensive consultation with implementers and end users it was determined that simplified dose forms were more useful for prescribing, dispensing and recording medicines. This also applies to including reference to route administration as part of the dose form; medicines can have multiple on label and off label routes of administration, and we are of the view this is clinical knowledge, and is up to the clinician to instruct the route a medicine is to be administered. For example fluoxetine dispersible tablets maybe be swallowed whole, but may also be dispersed in 100 mL of water and then the dispersion is swallowed; ceftriaxone 1000 mg injection can be administered via intramuscular injection or intravenous infusion.

The AMT avoids representing transformation of dose form, again we would consider this to be clinical knowledge that a clinician would be required to prepare a medicine prior to administration - ie; a powder for injection would imply that the powder would require reconstitution prior to injection, or a patient prescribed a dispersible tablet would receive instruction via prescription and dispensing label to disperse that tablet in some volume of water before drinking the suspension.

When incorporating new branded products, AMT assesses whether the product introduces a meaningful distinction compared to existing content. If it is materially different, it may be modelled separately; otherwise, the preference is to use the least specific dose form that remains accurate and useful.

Out of curiosity, is your approach mainly an ontological or clinical modelling accuracy question, or is there a particular use case for these interim dose forms? And are you able to share these use cases?
We have generally found that AMT users do not need this level of detail for common prescribing, dispensing and recording use cases, so it would be interesting to understand whether a different use case is putting more weight on these distinctions.

2 Likes
4

Hi Michael

My question is more modelling accuracy question and aims to understand why certain manufactured dose form are specified, for example with the term ‘concentrate’, while others are not, especially since retaining such specificity in some cases but not in others leads to the creation of multiple concepts representing the same medicinal product.
Based on your response, it appears that the way clinical drugs are described is both culturally influenced and linked to healthcare professionals’ practices, which in turn impacts the creation of clinical drugs in SNOMED CT.

5

Yes @Elisase, I think you really nail the problem with international concepts at the Clinical Drug level of specificity when you say:

it appears that the way clinical drugs are described is both culturally influenced and linked to healthcare professionals’ practices, which in turn impacts the creation of clinical drugs in SNOMED CT

That is exactly the issue. I would add that those cultural and professional-practice differences are themselves shaped by jurisdictional regulatory practices, subsidy and reimbursement models, and local implementation history. All of these affect the level of specificity that is needed at the Clinical Drug level to make their systems work.

To achieve alignment of Clinical Drug concepts in the International Edition, those deeper jurisdictional differences would need to be resolved. Compared with that, resolving the different modelling patterns and technical issues almost looks easy. Although those modelling differences are probably also a product of the same jurisdictional pressures.

This is why alignment at the Medicinal Product level, essentially active ingredient, is relatively straightforward. Medicinal Product Form, at least at a generalised form abstraction, is also reasonably tractable. But the step down from there to Clinical Drug opens a much larger can of worms.

At that level, researched through a number of different papers and experiments, jurisdictions often do not align. Drug extension builders usually do not have the flexibility to change or ignore their local regulatory, subsidy, reimbursement, or clinical obligations. Meanwhile, international content can only be modelled one way, which effectively means choosing one pattern as the winner at the Clinical Drug level.

That’s good for the winner, but for the other jurisdictions international Clinical Drug content then becomes something they necessarily inherit but need to avoid, work around, and explain to vendors and implementers as a distraction, rather than something that provides benefit.

1 Like
6

Thank you Michael, Elisabeth and Dion for raising and developing these points. This is a very relevant discussion, and it highlights an issue that goes beyond technical modelling choices.

The different approaches taken by drug extensions reflect not only ontological or modelling preferences, but also local clinical practice, regulatory requirements, reimbursement models, implementation history, and the needs of national systems. As Dion notes, this makes alignment at the Clinical Drug level particularly challenging, because the same level of specificity may be essential in one jurisdiction and unnecessary, or even problematic, in another.

For that reason, we think there is an important opportunity to make these differences more explicit and easier to understand. As part of the upcoming Drug Model Implementation Guide, we have decided to include an annex describing the different approaches used across drug extensions. The aim would be to document the modeling choices, their rationale, and the use cases they support, in order to support model improvements and new editorial advice.

We will be contacting the responsible teams in member countries to invite them to contribute short “drug extension profiles” describing their approach. These profiles could help extension builders, implementers, and vendors understand why particular modeling patterns have been adopted, where alignment is feasible, and where local requirements need to be preserved.

Hopefully, this will provide a practical way to share experience across countries, support better implementation decisions, and make the diversity of approaches more transparent within the SNOMED CT community.

2 Likes