Blog

Common Mistakes When Transcribing Using Speech Technology

In a typical busy hospital setting, some of the service providers try and opt to streamline the whole process of documentation of patient encounters, with the help of speech recognition software. This is done via the dictation of notes, that are subsequently saved in an electronic health record (EHR) system. However, this certainly does not mean that this otherwise excellent system is not prone to errors, especially during the actual process of transcription itself, that is quite prone to human errors, of both omission and commission.

Some of the more common speech transcription technology errors include the following:

  • Verbosity: Excessively added words
  • Pronunciation errors: Lack of knowledge of proper pronunciation
  • Multiple deletions: Stating and deleting words
  • Homonyms
  • Nonsense errors: Words being said seem to make no sense
  • Spelling errors, and last but not the least
  • Suffix or dictionary errors

Taken together, all of the above errors can easily create a lot of confusion when a person is taking notes and therefore, it would be well to ensure that proper care is taken when one is transcribing notes, onto a digital format via voice recognition software.

Apart from that, people also speak very quickly, which can create problems for the software when it is transcribing the data from voice to written speech. Furthermore, many doctors as well as their support staff do not take into consideration the fact that there is usually a lot of background noise in just about all hospital environments. While people working in such noisy places have acclimatized themselves to the overall ambient noise levels, and can easily write down their thoughts, the same does not hold true for voice recognition software. It can easily make a mistake and capture background voices, along with the speech that is actually being transcribed.

The crucial importance of double checking data

Such transcription errors are far more commonplace then they are assumed to be, especially in a real time environment, as stressful and noisy as a hospital. Here, such issues can easily crop up all over the place, due to the fact that many medical staffers rarely bother to go through their notes, before inking heir initials on the finished file.

Here it cannot be stressed enough that speech recognition errors can easily affect patient treatment. This is why the service providers using this technology should always be extra diligent with regard to spotting any issues that may potentially crop up during the process of documentation. Scanning the notes as soon as they are transcribed is a great way of ensuring that those troublesome errors are kept to a bare minimum. Apart from that, the final copy should also be proofread at least twice, in order to ensure the overall reliability of the document.

Ultimately, education, training and due diligence are all essential components in making sure that the number of such transcription related errors is kept as low as possible.