What is a map set in our world of medical terminologies?
A map set would relate concepts that are in two different code systems. A mapping would be established from a concept in the source terminology (“starting point”) to the target terminology (“ending point”) – thus the map has a direction, and you should not assume that mapping would be the same in the opposite direction.
The mapping relationship from the source concept to the target concept should be specified, as it may not be “is the same as”, which people tend to assume. The type of relationship would depend on the use case; for example, you may need to find the single closest map in the target terminology, so you can put the source concept in the same parent grouping. You should also be aware that mapping could be one-to-many (i.e., one source concept maps to multiple concepts in the target terminology), many-to-one (i.e., multiple source concepts map to the same concept in the target terminology), or one-to-one (i.e., one and only one source concept maps to one and only one concept in the target terminology). Each type of map has its usefulness in healthcare applications.
In a terminology database like the 3M HDD, which is concept based, we can represent map sets two different ways:
• Explicit mappings – between two concepts that are not clinically equivalent in meaning, the mapping is represented by a specified relationship from the source to the target, in the relationship table.
• Implicit mappings – between two concepts that have clinically equivalent meaning, their codes and descriptions (“representations”) would be associated with the same HDD concept, in the representation table.
Here is an example of explicit mapping:
The Canadian Classification of Health Interventions (CCI) codes are alphanumeric and up to a maximum of 10 characters with six discrete code “components”. This structure is similar to ICD-10-PCS but is not equivalent. As shown below, mappings could be one-to-one or one-to-many.
|CCI Code||CCI Description||ICD-10-PCS Code||ICD-10-PCS Description|
|1OD89LA||Excision total, gallbladder open approach without extraction of calculi cholecystectomy alone||0FT40ZZ||Resection of Gallbladder, Open Approach|
|2PQ71BA||Biopsy, urethra using endoscopic per orifice approach||0TBD8ZX||Excision of Urethra, Via Natural or Artificial Opening Endoscopic, Diagnostic|
|1TM93LAXXA||Amputation, elbow joint using skin graft (for closure of stump)||0X6C0ZZ||Detachment at Left Elbow Region, Open Approach|
|0X6B0ZZ||Detachment at Right Elbow Region, Open Approach|
|0HRBX73||Replacement of Right Upper Arm Skin with Autologous Tissue Substitute, Full Thickness, External Approach|
|0HRBX74||Replacement of Right Upper Arm Skin with Autologous Tissue Substitute, Partial Thickness, External Approach|
|0HRCX73||Replacement of Left Upper Arm Skin with Autologous Tissue Substitute, Full Thickness, External Approach|
|0HRCX74||Replacement of Left Upper Arm Skin with Autologous Tissue Substitute, Partial Thickness, External Approach|
Here is an example of implicit mapping:
Drug ingredients from SNOMED CT and RxNorm are mapped to the same HDD concept, indicating that they are considered clinically equivalent in meaning. Implicit mapping in the HDD is always one-to-one and the mapping relationship is always clinically equivalent, hence an explicit mapping specificity is not needed (would be redundant). This is also the only time that the map direction (who is source and who is target) doesn’t matter.