Auto-mapping of text-strings
A new member of the HealthTerm family is the Auto-Mapping module. Many healthcare providers are storing their information as free text – or local codes, which is a big challenge for integration, analytics as well as decision support.
HealthTerm makes it possible to take these strings and import them into the repository as a local code system. After this, you can now use all the HealthTerm modules for maintenance of this code system – including mapping, as well as all API’s for lookup.
The new Auto-Mapping module makes it easy to take local strings and match them through an advanced algorithm to a standard code system e.g. SNOMED CT, RxNORM or LOINC.
- The algorithm is intelligent and learns from previous mappings, as well as having specific filters for each standard code system to optimize the match. HealthTerm will find the best match for each term, but will suggest up-to 20 other matches – sorted after how well they match.
- The match algorithm is based on a combination of tokens, hierarchy distance, text distance and different search types and will use the structure of the standard code system to get the best match.
- The Auto-Mapping module is an integrated part of the standard mapping setup – which means that you can pass the suggested mappings through a workflow for qualification and have a full audit-log of the process.
- The Auto-Mapping module can be used as an API-service or as a full GUI implementation. All suggestions can be exported to Excel as either a best match – or all found matches (top 20).