AutoEncoder is easy to use as outlined in the steps below.
A background service detects new and changed verbatim terms in the source system and automatically feeds these into the AutoEncoder. Normally, this will happen within a couple of minutes.
Each verbatim term is assessed against the entries in the associated dictionary version, supplemented by records of prior history matches and any pre-loaded synonym lists. The assessment can include associated “reference items,” such as the country code for a medication. If a match is found for any of the sources, a “proposed match” is assigned to the verbatim term.
A coding user can either accept a proposed match or can perform a manual search. This manual search includes linguistic pattern matching. The text search can be combined with a search on any of the other dictionary columns.
The coding process may be followed by independent review and approval steps. A configuration option or elevated privileges for the coder may be used to bypass this step.
The coding results are automatically transmitted back to the source system, where they are stored as derived items. If the coding user subsequently decides to un-code a term, the coding results are blanked out in the source system and the term is sent back to the initial auto-code step. This means that the source system and the AutoEncoder are always synchronized with each other.