Using Speech Recognition Software Raises New Issues

posted by Kathy Everitt on Friday, October 05, 2018

Using Speech Recognition Software Raises New Issues

Today, there is increasing concern over the audibility of dictation and its impact on record accuracy.

Speech recognition technology (SRT) has two dimensions: front-end recognition and back-end recognition. 

Front-end speech recognition is when text is generated on-screen from dictation in real-time. This allows the healthcare provider to edit and finalize documents themselves. 

Back-end speech recognition is when completed dictations are automatically processed by the speech recognition server in the “background.” A transcriptionist receives a transcribed document and the original audio file. The transcriptionist’s role is to check the accuracy, completeness and format, rather than having to transcribe the entire report.

Both are meant to produce legible documents at a lower cost than a traditional transcription service. An additional benefit of SRT may be that it offers the opportunity to provide more detailed information to share with other providers in a more timely fashion.

When transcribed, both SRT forms should be held to the same standard as traditional transcription. Common SRT transcription errors include word omission, word substitution, verb tense, spelling, incomplete phrases, punctuation errors and age/gender mismatching. 

Regardless of the SRT used, make sure to have policies and procedures in place so documents are reviewed and edited consistently:

  • Consider (if using front-end technology) having a medical editor provide a quality check after the provider’s review. We are all guilty of seeing what we want to see on a document rather than what is really there. This second set of eyes may improve quality.
  • Determine accuracy standards for all documentation (handwritten, forms, checklists, transcription, free text, content)
  • Use disclaimers sparingly and only after they have been reviewed by your attorney. Disclaimers acknowledge errors in the content; however, they  may give the impression the provider is careless. Most importantly, disclaimers may not enhance the defense of the record in the event of a claim/suit.

For a quality and best practices tool kit related to documentation, visit http://www.ahdionline.org/page/qa

Speech recognition technology (SRT) has two dimensions: front-end recognition and back-end recognition. 

Front-end speech recognition is when text is generated on-screen from dictation in real-time. This allows the healthcare provider to edit and finalize documents themselves. 

Back-end speech recognition is when completed dictations are automatically processed by the speech recognition server in the “background.” A transcriptionist receives a transcribed document and the original audio file. The transcriptionist’s role is to check the accuracy, completeness and format, rather than having to transcribe the entire report.

Both are meant to produce legible documents at a lower cost than a traditional transcription service. An additional benefit of SRT may be that it offers the opportunity to provide more detailed information to share with other providers in a more timely fashion.

When transcribed, both SRT forms should be held to the same standard as traditional transcription. Common SRT transcription errors include word omission, word substitution, verb tense, spelling, incomplete phrases, punctuation errors and age/gender mismatching. 

Regardless of the SRT used, make sure to have policies and procedures in place so documents are reviewed and edited consistently:

  • Consider (if using front-end technology) having a medical editor provide a quality check after the provider’s review. We are all guilty of seeing what we want to see on a document rather than what is really there. This second set of eyes may improve quality.
  • Determine accuracy standards for all documentation (handwritten, forms, checklists, transcription, free text, content)
  • Use disclaimers sparingly and only after they have been reviewed by your attorney. Disclaimers acknowledge errors in the content; however, they  may give the impression the provider is careless. Most importantly, disclaimers may not enhance the defense of the record in the event of a claim/suit.

For a quality and best practices tool kit related to documentation, visit http://www.ahdionline.org/page/qa

Blog Author

Kathy Everitt

Senior Risk Management Consultant

Kathy brings with her more than 30 years of professional liabil...

Read More

BLOG CATEGORIES