This article explains the importance of picking a document type for dictation, the issues that can occur when using the wrong document type, and the use of default document types.
Drafts are produced from a speech model. The speech model is created from a speaker and document type paring.
In the screenshot above the Emdat Document Type “OpNote – Operative Report” is linked to the Nuance Document Type “Operative Note”. Operative Reports have a specific format. ASR learns this format from dictation and transcription processed during the model building phase.
When the clinician selects an Emdat Operative Report, ASR knows to use the Operative Note speech model to produce the draft. Corrections made to the draft update the Operative Note speech model to improve draft quality over time.
When a clinician doesn’t pick the appropriate document type, draft quality degrades because the wrong model is used. Example: the clinician picks the Operative Report document type but dictates a Letter. A draft is produced but has poorer quality because a Letter’s content and format is different from and Operative Report. The MT/MLS has to switch the document type to the Letter and then spend more time on corrections, decreasing productivity. Since the document typed is switched, the Letter model is used and receives the corrections. This is why it’s always a best practice for the clinician to pick a document type.
The example described what happens with an Operative Report document type, but this case applies to any Emdat Document type. EXCEPTION: When the Emdat Document Type is switched to the correct one, and both document types share the same Nuance Document Type draft quality will not be impacted because the same speech model is used.
Despite best practices, some clinicians prefer not choose a document type but the MTSO still wants to get the benefit of ASR. Depending on the clinician’s workflow, two scenarios exist. A) they have one document type on their device and choose not to select others. B) their device has no document type set at all.
In scenario A) jobs will go to ASR but will likely have poorer quality because the wrong speech model is used. In scenario B) the dictation will not go to ASR at all because no document type is indicated on the audio file. A document type must be indicated for the jobs to go to ASR.
For scenario A) the best strategy is to set the clinician’s device to the document type they dictate the majority of the time, their “default document type”. Example: if they dictate Consults 80% of the time and Operative Reports 20% of the time then at least 80% of the time the correct model will be used. For the other 20%, the MTSO will understand that those drafts produced will require more editing effort.
For scenario B) a default document type can be setup in the system. This guarantees the jobs will go to ASR, BUT they will still face the same quality issues as outlined in the scenario A strategy.
To set a default document type, go to Users > Edit the user, expand Dictating Clinician options and set the default document type there. It may be prudent to do this for scenario A as well in the event the clinician’s device is either misconfigured or a document type is not selected.