Page History

The purpose of the Substance Hierarchy redesign project is to eliminate errors and improve the consistency and usability of the substance hierarchy by reducing the overloaded subtype relationships and adding new attributes to capture this information correctly. The parts of the strategy involve:

1. Review and redesign of the Allergen class hierarchy

2. Increasing comprehensiveness of the model by adding relationships where necessary

3. Identification and removal of non-defining characteristics, outside of the allergen hierarchy

4. Enhancing and applying Editorial guidelines and rules to the descriptions of the substance concepts

5. Ongoing systematic inspection of each individual hierarchy

This project has been previously stalled due to resource shortages, in addition it was decided that further planning needed to be done with regards to how to approach Substances before executing any changes. The Substance plan has been completed however some points below may not have been in the execution scope.

SOW work - Draft -Substance Project Draft Outline

While past work on the project has focused on the pharmaceutical substances used in drug modelling, this phase of the redesign project includes all the substances and the multiple sub-hierarchies, including the drug and medicament substances, allergens, toxins and dietary substances.

• Improving the consistency of the modelling of the  substances and their descriptions
• Redesign representation of the allergen groupers and support allergen classes use cases
• Reduction of incorrect inferences causing issues throughout multiple SNOMED CT hierarchies
• Quality improvement, usability and accuracy of the Substance hierarchy and multiple other hierarchies throughout SNOMED CT.

• Allergen class redesign - document requirements and propose progression plan.

The ‘allergen class’ hierarchy is causing a number of quality issues for a number of reasons. Firstly, the relationships between allergen concepts and other substance concepts in the hierarchy are causing a number of incorrect inferences through inheritance. Secondly, the hierarchy is incomplete, further reducing the usefulness of the hierarchy. While there are over 7000 concepts in the “Drug and medicament” substance hierarchy, only half of these are covered by ‘Drug allergen’ (4700).

 2. Increasing the comprehensiveness of the model -document requirements and propose progression plan.

Past meetings and project groups have often raised the issue of the lack of relationships between certain substance concepts and the need to identify links between other concepts. For example, some substances which act prodrugs or metabolic precursors to other substances currently have no link, even though they may be closely related clinically. Another large issue that has been raised by the project group is the requirement for a relationship between the salts and the bases of a substance.

3. Identification and Removal of non-defining characteristics-document requirements and propose progression plan.

There are a number of concepts that describe substances based on non-defining extrinsic characteristics, for example, therapeutic roles. The inclusion of these results in many substances being incorrectly organised. The therapeutic purpose is something intrinsic to a product, and often associated with a specific dose or form.

4. Editorial Guidance/Rules-document requirements and propose progression plan.

There are currently few editorial guidelines which define the consistent naming conventions for substances in SNOMED CT, particularly in the grouper concepts in the hierarchy. Guidelines are also required to help determine when a grouper should be created and when a grouper definition is required. It is expected that this will be an ongoing process as required work throughout the whole redesign project.

5. Structure of the Overall Hierarchy/Systemic inspection-document requirements and propose progression plan.

As there is a large number and variety of sub hierarchies and content, the whole hierarchy would benefit from a manual inspection by members of the content team with knowledge of chemical, physiologic, therapeutic, preparation and manufacturing aspects of chemicals and devices, as well as a fundamental understanding of the principles of sound ontological design and description logics. 

Additional information -may not be inclusive to plan.

PSK

 1. will be deactivated by Allergies team

2. plan and progress Substance re-model in hand with Drugs

3. Therapeutic groupers are being dealt with in Drugs and relationships to Substances are treated accordingly.

4. Ed Guides have been started and progress as expected

5. Throughout the project a document is being written to document decisions taken in re-modelling of Substances and steps carried out in the process

1. A new Disposition hierarchy with new semantic tags, attribute types and type changes. Updates of modeling, definition status and terming with associated change in stated parents for Disposition Groupers (n ≈ 150) for existing descendants (n ≈ 9000)
2. Changes and updates of grouper hierarchy.

   • Disposition Groupers are updated, modelled and fully defined. They will be updated to proximal primitive and additional Disposition concepts and Disposition Groupers to be created.

   • Role Groupers have been reviewed. Some concepts have been inactivated and others replaced. 

   • Structure Groupers will be retained as primitive concepts. Additional Structural Groupers to be created.

   • Combined Groupers (Structure + Disposition) will be remodeled as we progress, the volume for this category is small.

   • Miscellaneous types of groupers will be addressed on a case by case basis

   • Modification groupers will be remodelled using role chain and DL features
3. The Editorial Guidelines are being updated following Project Group review in December 2017. Further updates to these guidelines are planned as work progresses

1. All changes in scope for July 2018 release are complete.
2. Tasks in scope for January 2019 are in progress

• Flattening of Substance hierarchy: Added/updated “Is modification” attribute where applicable: 633
• Disposition groupers:
  • Added new dispositions to model existing content: 41
  • Added new structural grouper concepts to model existing content: 61
  • Added/Updated modelling for existing concepts by adding "has disposition" attribute where applicable: 542
• Improve substance concepts:
• Updated modelling for existing concepts by adding missing structural parent(s): ~1500
• Added new substances (majority required for creation of new medicinal products): 162
• Improved substance concepts by aligning them to the International Nonproprietary Name: 380
• Primarily, but not limited Substance terms used in the Products hierarchy the Preferred Term in the US dialect has been made Acceptable in the GB dialect and vice versa.: 633

  • Substance concepts that went from P/N to P/A: 422

  • Substance concepts that went from P/N (or in 1 case P/A) to P/P: 221

• Improved substance concepts by Fixing odd symbols like "[" in substance terms: 30
• Improved substance concepts by updating case sensitivity based on editorial guidelines in the following areas: ~5000
  
  • Greek alphabetical terms in substance concepts have been made case insensitive
  • Single word substances incorrectly case sensitive amended
  • For substance terms that have a single letter - either a single lower case letter or a single upper case letter - the case sensitivity has been changed so now correctly reflects this
  • Substance terms containing proper names now reflect this capitalization requirement, and quality checks have been performed on the substance hierarchy based on the file used for automated spell-checking and automated case sensitivity checking. A decision was taken to make the substance hierarchy the 'source of truth' for other hierarchies within SNOMED CT.
  • Rule established that numbers at the beginning of substance terms can be ignored for case significance, and the case sensitivity assignment now starts with the first letter, including batch amendment of incorrect capitalization in terms that include numbers.
• Fixed issues identified in the QA reports:
  
  • Fixed issues with duplicate descriptions: 4
  • Fixed issues with missing historical associations: 4
  • Fixed issues with spelling mistakes: 25
• Changes in the substance hierarchy as the result of Allergy project requirements:
• Inactivate top level groupers/Allergen classes in substance hierarchy by reassigning their children to appropriate structural and/or functional parents (> 100 stated children and 100s of inferred children)
• 414052000 | Drug allergen by structure |
• 414058001 | Drug pseudoallergen by function |
• 414057006 | Drug pseudoallergen |
• 414053005 | Drug allergen or pseudoallergen |
• Inactivate “Allergy to X” concepts where X is a modification of a base substance and replace it with Allergy to the base substance (~10s)
• Update Editorial Guidelines

All changes in scope for January 2019 release are complete.

Tasks in scope for July 2019 to be reviewed and priorotized