Healthcare transcription and speech Recognition

Is speech recognition the new way out for medical transcription requirements? Voice recognition or speech recognition is one of the new growths in the medical transcription field. It has been flaunted as the solution for a health professional’s documentation needs.

How does it pan out? The doctor is offered a digital platform by the voice recognition company where the healthcare professional can call in and dictate his patient notes at the end of the day. The software converts it to typed form, which is made available to the doctor to edit and correct as needed

Seems simple enough. But is it as simple as that? Apparently not! Speech recognition requires to be evaluated on all the standards used while deciding upon a hospital’s medical transcription company.

Time saved: When medical transcription needs are contracted out one of the most important benefits is the time conservation aspect. Time is a valuable commodity for healthcare professionals and they would like to utilize it for providing excellent care to their patients. But when voice recognition software is used the doctor has to spend more time on it in the following areas:

Preparation time: speech recognition software needs the physician to personally ‘develop’ the software on the speech example, accent, standard terms and various nuances of his/her speech pattern so that the software can ‘learn’ it and acclimatize in accordance with. This could involve a substantial time commitment from the doctor operating it, which the physician may ill afford to make.

Time for dictation: speech recognition software requires the healthcare professional to change his style of dictation to include punctuations, grammar, spellings, detailed terminology, paragraph and report start and ending etc., as opposed to the free flow form of dictations that the physicians are used to.

Time for proofreading: Unlike the traditional method of medical transcription where the patient transcripts are subject to thorough proof reading and quality verifications, here the physician needs to spend considerable time editing and correcting before signing off on the reports
.
Accuracy: The accuracy levels of the finished reports is of utmost importance as it affects care of the patients at every level as well as forming the basis for all further documentation needs. The accuracy level from voice recognition software is found to be about 60%-65%. That means correcting: and editing the transcripts to make up for the residual 35%-40% quality.

Costs: The price aspect for voice recognition software will fluctuate based on the number of people using and the size of the organization. It will also have to factor in additional expenses needed for correcting the reports to an acceptable level. One may find that it is costing more per line than was quoted in the beginning.

Time to become adept: When healthcare providers seek out for solutions for transcribing their patient medical records they would like to have services that are competent and exact from the very beginning. In the case of speech recognition software there is a definite time to become adept that affects the correctness of finished transcripts till the software ‘learns’ to follow a particular physician’s talking patterns and even then the accuracy levels are only 60%-65%.

Adaptability: When new healthcare professionals are added to the health facility the speech recognition software has to be taught further to receive the speech nuances of the new joiners. There is no agility provided by voice recognition software for this.


In conclusion we can definitely say that voice recognition software will never be able to absolutely replace manual medical transcription. The important factor to be evaluated here is that patient transcripts are created of human beings, dictated by human beings and therefore needs another human being to transcribe it. Medical transcriptionists will have a valuable role to play at least as editors and proofreaders.

Healthcare transcription requires a certain amount of training, subjective judgment and understanding the context to be exact and relevant, which will not be possible to achieve completely through software, no matter how expertly created.

For the transcription requirements of a healthcare facility one should look for a vendor that delivers dedication on quality, turnaround time, reasonable costing and HIPAA compliance