« »

Friday, December 26, 2014

How voice recognition technology is changing medical transcription

Medical transcription software
 Evolving medical transcription software improves record keeping
By Mary Lynn

As hospitals update their information processing systems, voice recognition software is one of the new technologies being implemented. This technology makes it possible to feed voice commands to a computer and then use software to automatically transcribe audio into text. In a hospital environment, doctors can use voice recognition applications to dictate patient notes and seamlessly create written documents without having to type anything. The prospects of including such automations into the patient record management systems of hospitals are indeed significant.

With speech recognition software, hospitals can process patient details more efficiently and reduce overhead costs that go towards paying salaries of transcription staff. But, what does this mean to audio typists and the role they play? Is voice recognition replacing manual audio typists? How feasible is this technology in the first place? This post will discuss these issues in order to paint a clear picture of just how voice recognition technology has changed the medical transcription industry.

The evolving role of MTs


It has been said time and again that speech recognition applications will wipe medical transcriptionists (MTs) to oblivion. That notion though is far from the truth. Even as the medical transcription industry continues to evolve with new technology, the demand for qualified audio typists is very much alive.  Although advances in technology have given doctors smart voice programs to automate record entry tasks, these applications aren’t as dependable as the human ear. Because of the complexity of natural language, most voice recognition apps are not accurate when it comes to transforming text into audio. These applications create transcripts that are more of a draft than a final document. Much of this has to do with how they operate, which makes it difficult to capture every word accurately in an audio file.

Typical speech recognition software usually stores a database of words that correspond to voice signals. These applications then try to match words with voice data fed to a computer. Due to differences in accents, homophones in language and the complicated nature of how people speak, errors are inevitable in any software generated transcript. In a medical report, these errors can mean the difference between life and death. This is why medical transcriptionists still need to listen to medical related recordings and proof transcripts for factual, grammatical and spelling errors. This therefore means that as hospitals continue to implement speech recognition technology, medical transcriptionists have taken more of an editorial role.

Efficiency in transcribing dictations


In spite of the limitations that voice recognition programs have, many hospitals benefited a lot by using them. With an accuracy of between 60 to 90 percent, speech recognition applications can indeed add an element of efficiency into the patient record entry process. These applications also do an excellent job with dictations that have only one speaker. Such recordings are easy for the voice recognition software to decipher. In most cases, these applications learn voices, hence increasing accuracy after getting used to a particular voice.

As technology continues to advance, perhaps voice recognition technology will make it possible to transcribe complex audio recordings with flawless accuracy.


About the author: Mary Lynn is a blogger based in Aberdeen who loves to write about technology and its impact in various careers. If you want to learn more about transcription and where Lynn outsources her personal audio typing needs, click here.

Image: Dr. Harry Gouvas, "A Medical Software International Logo"; CC BY-SA 3.0