Introduction

The SNOMED CT terminology provides a common language that enables a consistent way of indexing, storing, retrieving, and aggregating clinical data across specialties and sites of care. The International Health Terminology Standards Development Organisation (IHTSDO^®) maintains the SNOMED CT technical design, the content architecture, the SNOMED CT content (including the concepts table, the descriptions table, the relationships table, a history table and mappings), and related technical documentation. 

This document is intended to give a brief background to the Mapping Conventions used on the SNOMED CT International/ICH (International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use) work through the EU funded WEB RADR 2 project and the resulting Release package and the resulting SNOMED CT/MedDRA maps, which are being published as an Alpha release at this time with intention to publish a production release in April 2021.  It is not a detailed technical document of SNOMED CT, MedDRA or the SNOMED CT International/ MedDRA release. Nor does it seek to provide an editorial policy for this content though does provide details of the Mapping conventions applied.

This ALPHA release package is distributed for evaluation purposes only.  It must not be used in production clinical systems or in clinical settings, or distributed to Affiliate Licensees or any third parties.   

Audience for this document

This document should be read by all those (SNOMED International National Release Centers, vendors of electronic health records, terminology developers, Regulators and Pharmaceutical Industry) with an interest in the usage of this content in SNOMED CT and its linkage with MedDRA i.e. the SNOMED CT to MedDRA and MedDRA to SNOMED CT maps based on a priority set of MedDRA. 

Background

In 2018, SNOMED International and ICH joined the WEB-RADR 2 project, funded by Innovative Medicines Initiative (IMI). SNOMED International and ICH lead the effort to develop 2 maps between SNOMED CT and MedDRA. A specific work package was devoted to this purpose and included participants from regulatory authorities and industry as well as SNOMED International and MedDRA MSSO. The project delivered the following:

• 2 maps (MedDRA to SNOMED CT and SNOMED CT to MedDRA) of frequently used pharmacovigilance MedDRA terms that were identified from databases from the UK’s Medicines & Healthcare products Regulatory Agency (MHRA) and the European Medicines Agency (EMA). This subset represents the high value terms to support interoperability between the terminologies.
• A set of Mapping conventions which were drafted as part of a pilot set and updated for the main body of work.
• A small number of content additions to SNOMED CT in support of the MedDRA high value terms as defined by scope and editorial rules of SNOMED CT.

The Use cases of the project were defined as:

SNOMED CT to MedDRA map:

Key pharmacovigilance concepts when coded in SNOMED CT in an electronic health record could be converted to MedDRA for the purpose of adverse event reporting to regulatory authorities or for the purposes of epidemiological research. 

The mapping work was undertaken using dual independent review by mappers who were Consortium members of the WEB RADR 2 project and experts in either SNOMED CT or MedDRA. Quality Assurance, based on the mapping conventions, was managed as part of the process by SNOMED International and MedDRA MSSO (on behalf of ICH) and technical assurance reports provided weekly using the Mapping tool used by all participants. A big thank you is offered to all contributors to the Maps. The Mapping Conventions used are provided for information as part of this package.

In parallel to the mapping work, the operational aspects have been worked on to agree activities and processes to enable the 2 Maps to move in to Production in April 2021 and from there to be maintained and updated according to changes within the 2 terminologies and Use Case based requirements for additions to the Maps - thus meeting one of the key requirements of IMI to produce products in the project that are sustainable over time.

At this time, SNOMED International and ICH are seeking feedback on the Alpha version of the 2 Maps, and a Test/Review protocol has been produced as part of this package which indicates some of the areas on which feedback is invited. The Test/Review period ends on 30 September 2020, after which work will be undertaken to update the Maps in line with January 2021 SNOMED CT International Release and MedDRA September 2020 release. A report of the feedback will be made available. Please note that the Maps will NOT be extended beyond current scope for the Production release. The Production release in April 2021 will be available to licensed users of either SNOMED CT of MedDRA.

MedDRA to SNOMED CT map:

Clinical care decision support: Terms coded in MedDRA representing adverse events, warnings, and other regulatory information in product labeling could be converted into SNOMED CT so that the information is integrated into the electronic health record or decision support system to be visible to healthcare professionals when prescribing, dispensing, or administering the product to patients.

Use Cases of the Maps 

A major goal of the WEB-RADR 2 project is to use the enhanced functionality of the mobile application to facilitate exchange of data between regulatory databases (which use MedDRA) and healthcare databases/electronic health records (which use SNOMED CT). Two maps are therefore being developed (from MedDRA to SNOMED CT and SNOMED CT to MedDRA) to support seamless data exchange within the application platform. The starting sub-set of ~7,400 terms to be mapped in the project define a set of key pharmacovigilance terms that need to be linked to their counterparts in either terminology.

In one use case, these key pharmacovigilance concepts when coded in SNOMED CT in an electronic health record could be converted to MedDRA for the purpose of adverse event reporting to regulatory authorities or for the purposes of epidemiological research. In the opposite direction, these same key terms coded in MedDRA representing adverse events, warnings, and other regulatory information could be converted into SNOMED CT so that the information is available in the patient’s record to aid in clinical decision-making. 

Timeline for Creation of the Maps 

Test mapping (760 terms, February-March 2019)

› MedDRA v21.1

› SNOMED CT version Jan 2019 International edition

» Completion of mapping (1 May 2019 onwards, confirmed after test)

› MedDRA v21.1
› SNOMED CT version Jan 2019 International edition

» Impact assessment of July 2019 release of SNOMED CT and MedDRA v22.0 and v22.1 (after completion of the map)

Post-mapping note: in finalizing the mapping, an impact assessment was performed to determine which MedDRA terms had been made non-current and which SNOMED CT concepts had been made inactive. These terms were then removed from the mapping. No other changes were made to the mapping based on terminology version updates, i.e., any terms that had been proposed and added to SNOMED CT or any medically better matches to a mapped term were not included in the final maps. 

The final maps are based on:

› MedDRAv21.1

› SNOMED CT version Jan 2019 International edition

Any terms that are non-current (MedDRA) or inactive (SNOMED CT) as of October 2019 are excluded.

Note that it proved not possible to map all of the starting sub-set of ~7,400 terms given the different scopes of the two terminologies and the way they are structured, i.e., some terms were unmappable. Other terms were identified as potentially mappable and will be considered for possible addition to SNOMED CT in a future release. In finalizing the maps, other actions were taken that impacted the number of terms; these include adding missing US and UK spelling variants to the MedDRA to SNOMED CT map and removing unqualified test name terms from both maps. These actions are documented in Post-Mapping Notes below. Overall, the final maps produced in this project have decreased numbers of terms compared to the starting sub-set. 

General Mapping Guidance 

MedDRA hierarchies 

MedDRA groups its terms in a five level hierarchy. The Preferred Term (PT) level represents single medical concepts and the Lowest Level Term (LLT) level represents synonyms, lexical variants, and sub-elements. 

Determining possible equivalencies 

Check SNOMED CT concepts and MedDRA terms against hierarchy placement to determine if concepts/terms are equivalent.

Example

• ›  VitaminD. Direct lexical match but is a test name in MedDRA and a substance in SNOMED CT

• ›  In this case, LLT Vitamin D should be mapped instead to SNOMED CT Vitamin D measurement (procedure) 

Mapping to an exact conceptual match 

Mapping in both directions is to an exact conceptual match for the concept/term. Such mappings will all be flagged with the mapping parameter (attribute/relation) “Exact Match”. This is defined as the source terminology concept/term and the target terminology concept/term having the same conceptual medical meaning.

Examples

• ›  LLT Permanent cardiac pacemaker insertion maps to SNOMED CT Implantation of cardiac pacemaker (procedure)

• ›  SNOMED CT Implantation of cardiac pacemaker maps to LLT Cardiac pacemaker insertion

• ›  LLT Emotional lability maps to SNOMED CT Mood swings

• ›  SNOMED CT Mood swings maps to LLT Mood swings

Note that while Emotional lability is a synonym of Mood swings in SNOMED CT, in MedDRA, LLT Emotional lability is under PT Affect lability (HLT Affect alterations NEC) and LLT Mood swings is under PT Mood swings (HLT Fluctuating mood symptoms). Both terms are under HLGT Mood disorders and disturbances NEC. The two terminologies use different editorial guidance for their organisation and in some cases such as this one where the structure of SOC Psychiatric disorders is based on DSM-5, closely related terms may be in different parts of the hierarchy. For the purposes of the mapping, a pragmatic approach will be taken and concepts/terms will either be exact conceptual matches or unmappable.

Unmappable Concepts / Terms

Any concepts/terms that are not an Exact Match will be flagged as “Unmappable in SNOMED CT” or “Unmappable in MedDRA” as appropriate. This will identify relevant concepts in either terminology that might be missing and are required to provide a more complete mapping. The addition of any new content will then be discussed by the relevant terminology organisation.

Examples

• ›  LLT Non-ischemic cardiomyopathy and SNOMED CT Nonischemic congestive cardiomyopathy are not an exact match because the SNOMED CT concept has the additional concept of “congestive”. This would be flagged as “Unmappable in MedDRA” and a more specific term, Non- ischemic congestive cardiomyopathy, would be considered for addition to MedDRA.

• ›  LLT Prostate cancer stage IV and SNOMED CT Carcinoma of prostate are not an exact match since the stage is not included in the SNOMED CT concept. This would be flagged as “Unmappable in SNOMED CT” and a more specific term, Prostate cancer stage IV, would be considered for addition to SNOMED CT. (See also Principle 8 for conventions for tumour types and stages). 

MedDRA LLT to SNOMED CT FSN

The MedDRA to SNOMED CT map will be undertaken first (LLT to FSN). That same SNOMED CT concept (FSN) will then be mapped in the reverse direction to the equivalent MedDRA LLT, aiming for the same semantic match.

Mapping multiple LLTs to a single FSN

In many instances, the LLT mapped from MedDRA to SNOMED CT will be the same as the LLT when mapped in the reverse direction from SNOMED CT to MedDRA, i.e., LLT 1 to FSN; FSN to LLT 1. In others, the LLT mapped from MedDRA to SNOMED CT will differ from the resulting LLT when mapped in the reverse direction, i.e., LLT 1 to FSN; FSN to LLT 2. This occurs because the two terminologies differ with respect to lexical variants, spellings, etc. However, the clinical meaning of the term/concept should always be the same in both directions. See Principles 1 and 2 for specific examples.

While more than one LLT can map to a single FSN, in the reverse direction going from SNOMED CT to MedDRA, this will always be a 1:1 map. This supports the use case of using SNOMED CT EHR data to report Adverse Events without double counting.

Map active content only

Map active SNOMED CT concepts and current MedDRA LLTs only

Mapping domains

Typically the map will use (finding/disorders), (event), (procedure), and (situation with explicit context) concepts in SNOMED CT. However, there may be valid exceptions.

Example

› LLT Blood pressure systolic maps to FSN Systolic arterial pressure (observable entity).

The map will not use (substance) concepts in SNOMED CT since the names of drugs and other substances are out of scope of MedDRA. 

Specific Mapping Conventions

Principle 1. Map MedDRA LLT to SNOMED CT concept, finding the same semantic match.

Use Fully Specified Name (FSN) to confirm meaning. The same applies to the SNOMED CT concept to MedDRA LLT map.

Example

› MedDRA LLT Rhabdomyolysis maps to SNOMED CT Rhabdomyolysis (disorder)

- Do not map to SNOMED CT Rhabdomyolysis (morphologic abnormality) as this concept refers to a pathological related finding rather than a clinical one 

› SNOMED CT Rhabdomyolysis (disorder) maps to LLT Rhabdomyolysis 

Note that in this example, the starting LLT (Rhabdomyolysis) in the MedDRA to SNOMED CT map is the same as the LLT (Rhabdomyolysis) in the reverse SNOMED CT to MedDRA map. i.e., LLT 1 to FSN and FSN to LLT 1. 

Principle 2. When identifying maps, only use the synonyms in SNOMED CT to inform choice of the FSN. 

The tool will not permit mapping to synonyms.

Example

› MedDRA LLT Somnolence maps to SNOMED CT Drowsy (finding)
› SNOMED CT Somnolence is a synonym of Drowsy (finding) and cannot be used to map.
› Other LLTs under PT Somnolence such as LLT Sleepiness, LLT Sleepy, and LLT Daytime sleepiness are also included in the list of terms to map and these will also map to SNOMED CT Drowsy (finding), i.e., a many to one map. 

› In the reverse direction, SNOMED CT Drowsy maps to MedDRA LLT Drowsiness(1:1map) “Drowsy” is not in MedDRA, only LLT Drowsy on awakening. MedDRA LLT Drowsiness is closest match to SNOMED CT Drowsy. 

Note that in this example, the starting LLT (Somnolence) in the MedDRA to SNOMED CT map differs from the LLT (Drowsiness) in the reverse SNOMED CT to MedDRA map. i.e., LLT 1 to FSN (Drowsy) and FSN to LLT 2. This results from finding the closest match to SNOMED CT Drowsy which is LLT Drowsiness in MedDRA. All of the maps represent the same medical meaning however: LLT Somnolence and LLT Drowsiness both are under PT Somnolence and they are represented in SNOMED CT as FSN Drowsy (with its synonyms including Somnolence, Drowsiness, Sleepiness, etc.) 

Principle 3. Not Otherwise Specified (NOS) and Unspecified terms in MedDRA

• »  In the MedDRA to SNOMED CT map, map NOS and unspecified LLTs to the unqualified SNOMED CT concept, i.e., without any further classification

  Example

  • ›  LLT Pain NOS maps to SNOMED CT Pain (finding)

  • ›  LLT Non-auto immune hemolytic anemia, unspecified maps to SNOMED CT Non-auto immune hemolytic anemia (disorder)

    
    

• »  NOS and unspecified concepts will not be added to SNOMED CT

• »  In the SNOMED CT to MedDRA map, the SNOMED CT FSN is mapped to the unqualified LLT, i.e., do not map to NOS or unspecified LLTs in this direction

  Example

  • ›  SNOMED CT Pain (finding) maps to LLT Pain

  • ›  SNOMED CT Non-autoimmune hemolytic anemia maps to LLT Non-autoimmune hemolytic anemia

Principle 4. UK English and US English variants

For MedDRA to SNOMED CT map:

• »  Map both US and UK spelling variants in MedDRA to the SNOMED CT concept, i.e. the FSN (US and

  UK spellings (descriptions) are synonyms and can’t be used to map)

  Example

  › LLT Edema and LLT Oedema map to SNOMED CT Edema (cardinality of 1:1)

  
  

• »  If the MedDRA source data set contains only either the UK or US spelling of an LLT, the MSSO will

  identify and review the equivalent alternate missing spellings for possible addition to the mapping.

  Post-mapping note: in finalizing the mapping, all “missing” spelling variants in the MedDRA source data were identified and mapped to the appropriate FSN (US spelling). 

For SNOMED CT to MedDRA map:

Example

› FSNEdemawillmaptoLLTEdema(1:1cardinality) 

that clinical records using SNOMED CT use either UK or US spellings. When SNOMED CT is implemented in EHRs, a language subset (UK or US English) will be implemented by the system depending on the country of location. The end user will thus view the US or UK synonyms but these are represented by the unambiguous SNOMED CT concept which uses the US spelling.

Every PT in MedDRA has an LLT that is identical to it and shares the same code. In MedDRA, UK English spelling is used at the PT level and above; US spellings are only represented at the LLT level. Analysis is performed at the PT level. 

In the use case of taking SNOMED CT EHR data and converting it to MedDRA to report or count adverse events (AEs), one needs to avoid double counting. The 1:1 cardinality from SNOMED CT (US spelling) to MedDRA (US spelling) would ensure that events are only counted once in MedDRA. Maps are generated based on SNOMED CT concepts; whether the EHR uses the SNOMED CT US spelling or the UK spelling, both would map via the FSN to the same single term in MedDRA.

Principle 5. Combination terms and infection/body site

» Map MedDRA LLT to an equivalent SNOMED CT combination term if available

Examples

› LLT Dementia due to Parkinson’s disease maps to SNOMED CT Dementia due to Parkinson’s disease

› LLT Escherichia urinary tract infection maps to SNOMED CT Escherichia coli urinary tract infection

» Similar principles apply in the reverse SNOMED CT to MedDRA direction
» If an equivalent combination term is not available in either terminology, flag as Unmappable for discussion and potential addition 

Principle 6. Test names and test results

» MedDRA test name term concepts will generally be found in (procedure) concepts in SNOMED CT

Examples

• MedDRA LLT Cholesterol maps to SNOMED CT Cholesterol measurement (procedure), not SNOMED CT Cholesterol (substance)
  
  

» MedDRA test results are typically found in SNOMED CT (finding)

Examples

• MedDRA LLT Blood glucose increased maps to SNOMED CT Increased glucose level (finding)

• SNOMED CT Increased glucose level maps to MedDRA LLT Glucose increased (SOC Investigations) 

MedDRA has important exceptions for the following high/low, increased/decreased terms which are considered to be disorders rather than investigations based on the common use of these terms:

› LLT Blood pressure high (PTHypertension)
› LLT Blood pressure low (PTHypotension)
› LLT High temperature (PTPyrexia)
› LLT/PT Sputum increased (HLT Coughing and associated symptoms)

› LLT/PT Sputum decreased (HLT Coughing and associated symptoms)

The corresponding SNOMED CT concepts will also therefore be disorders.

Examples

› LLT Blood pressure high maps to SNOMED CT Hypertensive disorder, systemic arterial (disorder)

› LLT Blood pressure low maps to SNOMED CT Low blood pressure (disorder)

Post-mapping note: although unqualified test name terms were included in the MedDRA source data, it proved challenging to map many of them to SNOMED CT. In addition, since these terms only convey that a test was performed, they were considered not to be of value from a pharmacovigilance or clinical information perspective. Therefore, unqualified test name terms were removed from the final mapping. 

Principle 7. Specimen type

• »  If the specimen type is not specified in the source concept/term, it should be mapped to the concept/term without the specimen type, if available

• »  If the specimen type is not specified in the source concept/term and the concept/term without the specimen type is not available in the target terminology, it is acceptable to default to blood or the most common specimen type for that particular test

• »  The default specimen type will be serum, not plasma, if blood is specified. Try to preserve the specimen type whenever possible.

  Examples

  • ›  LLT Lactate dehydrogenase maps to SNOMED CT Lactate dehydrogenase level (not to SNOMED CT Plasma lactate dehydrogenase level)

  • ›  LLT Drug level increased maps to SNOMED CT Blood drug level high (there is no drug level high concept without the specimen type available so it is acceptable to default to blood)

  • ›  LLT Lactate dehydrogenase increased maps to SNOMED CT Serum lactate dehydrogenase level elevated (concepts for the increased/elevated term without the specimen type, or specifying blood or plasma are not available so it is acceptable to map to serum in this case)

  • ›  LLT Blood creatinine maps to SNOMED CT Creatinine measurement, serum 

Principle 8. Tumour types and stages

• »  If SNOMED CT does not contain an FSN with both the histopathologic type and site of the tumour, the MedDRA term should be flagged as “Unmappable in SNOMED CT” and reviewed for possible addition.

  Example

  › LLT Renal granular cell carcinoma does not have a corresponding FSN within SNOMED CT and should be marked Unmappable.

  
  

• »  SNOMED CT does not contain all stages for specific tumour types. If the tumour stage concept is not available in SNOMED CT for a particular stage term in MedDRA, the term should be flagged as “Unmappable in SNOMED CT” and reviewed for possible addition.

  Example

  › LLT Breast cancer stage III does not have a corresponding FSN within SNOMED CT and should be marked Unmappable.

  
  

• »  When a “recurrent” cancer term is not available in SNOMED CT then the term should be mapped to the primary cancer.

  Example

  › LLT Non-small cell lung cancer recurrent maps to FSN Non-small cell lung cancer 

  
  

• »  When a MedDRA term specifies cell type and site, in addition to staging, try to preserve all three aspects of the concept in the SNOMED CT map.

   Examples

   ›  LLT Lung adeno carcinoma metastatic maps to FSN Adeno carcinoma of lung, stage IV

   ›  LLT Adeno carcinoma of colon stage IV is unmappable due to the lack of atermin SNOMED CT that captures Adenocarcinoma + Colon + stage IV/4/metastatic, including the “metastatic from” form referenced below.

The exceptions to this principle are the stage 4 or metastatic cancers.

• »  For metastatic or stage 4 cancers, when an exact match is not available, “[Primary site] metastatic cancer” LLTs map to the similar “Metastasis from [primary site]” FSN

  Example

  › LLT Lung cancer metastatic maps to FSN Metastasis from malignant neoplasm of lung

  
  

• »  Metastatic primary site terms are synonymous with stage IV/stage 4 if the metastatic term is not available

  Example

  › LLT Lung adeno carcinoma metastatic maps to FSN Adeno carcinoma of lung, stage IV

  
  

• »  Close attention should be given to the “secondary from” and “metastatic to” terminology.“ Metastases [organ]” LLTs map to FSNs of “Secondary malignant neoplasm of [organ]”

  Example

  › LLT Metastases to spine maps to FSN Secondary malignant neoplasm of vertebral column 

Principle 9. Natural word order

» When mapping from SNOMED CT to MedDRA and there are several lexical variant LLTs, the one that uses natural word order should generally be chosen. However, because of the granularity of MedDRA, there may be several LLTs under a PT that are all equally acceptable. If the map specialists happen to choose different ones, the map lead will mark them as not in conflict and select one LLT as the mapped term.

Examples

• ›  SNOMED CT Hyperpigmentation of skin maps to LLT Skin hyperpigmentation (natural word order) rather than LLT Hyperpigmentation skin

• ›  SNOMED CT Collapse of vertebra could map to LLT Collapsed vertebra or LLT Vertebral collapse. Both use natural word order and are under PT Spinal compression fracture. 

Principle 10. Punctuation marks

SNOMED CT concepts typically do not include punctuation marks such as hyphens or slashes and will be mapped to the equivalent MedDRA term that also does not include these marks. If there is no MedDRA term without punctuation marks available, the SNOMED CT term will be mapped to the equivalent MedDRA term with punctuation marks. New terms without punctuation marks will be considered for potential addition to MedDRA.

Example

› ForSNOMEDCTAttentiondeficithyperactivitydisorderthereareonlypunctuatedLLTs:LLT Attention deficit/hyperactivity disorder and LLT Attention deficit-hyperactivity disorder

Principle 11. Increased/high and decreased/low qualifiers

In both MedDRA and SNOMED CT, qualifiers for investigation results such as increased/high/elevated and decreased/low are generally used synonymously and can be used to map in both directions.

In both terminologies, the clinical meaning of the “increased” qualifiers can refer to an increase from a previous baseline or an increase above the normal range for that parameter. Similar considerations apply to concepts/terms that are “decreased”. Since it is not known from the report which clinical meaning applies unless one reviews the actual laboratory data, the qualified investigation terms in each terminology will be mapped to their equivalent counterparts. In practice, it is likely that “increased” and “decreased” are generally used by reporters to refer to changes above and below the normal reference ranges (see Increased bilirubin level example below for how this is sometimes reflected in the SNOMED CT hierarchy). However, for the purpose of this bi-directional mapping, the concepts/terms will be mapped strictly to their equivalents since the underlying clinical meaning is not known on an individual case basis.

Example

› LLT Bilirubin elevated maps to SNOMED CT Increased bilirubin level (finding). Note that the interpretation is “Above reference range”. 

Example

› LLT Blood pressure systolic increased maps to SNOMED CT Increased systolic arterial pressure (finding). Note that in this case the interpretation is “Increased” but it does not state if this is above the reference range.