Is Speech Recognition for Medical Transcription a Good Option?

Note: This article was originally published in 2020 and may not reflect the latest advancements in how AI is used to automate medical documentation. For more up-to-date insights, check out this post on how ambient AI is transforming the medical documentation landscape.

— Archived content begins below

‍

Is speech recognition for medical transcription a good option for care providers?

New technology may suggest that speech recognition is the future of medical documentation. In a way, it is, but probably not in the way you think.

What is speech recognition for medical transcription?

In past weeks we’ve discussed the more traditional methods of how healthcare providers utilize medical transcription. Most commonly, these methods involve some form of dictation recording, which is followed by a medical transcriptionist completing the medical documentation and SOAP notes by transcribing the dictation. In recent years, however, strides have been made in the field of medical transcription. Now, there are a handful of solutions for healthcare providers that rely on speech recognition to produce medical transcription.

What is speech recognition for medical transcription and how does it work?

Rather than dictating patient information and waiting for it to be transcribed later by a medical transcriptionist, speech recognition for medical transcription works in real time, totally removing the need for a medical transcriptionist or an outsourced medical transcription company. This advanced speech recognition software uses artificial intelligence and natural language processing to capture the dictated elements of a patient encounter and transcribe them into a medical note. By relying on explicit vocal prompting, this speech recognition software can capture the information and apply it into the medical note quickly and accurately. 

Benefits of using speech recognition software for medical transcription

The biggest benefit to using speech recognition software to transcribe medical notes is that it happens in real time. Clinicians don’t need to endure long turnaround times to receive their notes, or pay more to have their transcriptions expedited, they can simply turn on the recording device and start dictating and documenting right away.

The other big benefit is cost, speech recognition technology is often cheaper than hiring a third party medical transcriptionist. While many of the popular providers charge a monthly subscription or annual license fee, the totals are often still far less than hiring a medical transcriptionist. For comparison, in-house transcriptionists can make as much as $50k per year in some cases, and outsourced transcription services can also be very costly.*

*Most outsourced transcription services have more complicated billing systems, often charging by the line, word, minute, or keystroke, so annual cost totals can be difficult to estimate as they vary by provider, note style, and frequency of use.

Drawbacks of speech recognition software leveraged for medical transcription

Despite speech recognition for medical transcription eliminating lengthy note turnaround times, the reality is that the technology is still very time consuming. This is due in large part to the fact that in order to use speech recognition software, a provider must dictate every element of their note as if they were typing it. This means that every minute punctuation detail, section title and data label must be logged vocally in order to appear in the medical note. Speech recognition software for medical transcription, despite using AI, has no predictive capability when it comes to deciphering or understanding context. It’s a highly intelligent system, but still requires a great deal of seemingly analog input. So much so that some clinicians often find themselves spending an equal or greater amount of time on their clinical documentation. Rather than typing notes, they are simply dictating them for a program to type instead.

The other under-discussed issue with using speech recognition for medical transcription and, truthfully, medical transcription in general, is that it doesn’t reduce the need for deep information recall. In other words, opting for a clinical documentation tool that can only be used after a patient leaves the exam room means providers must rely on detailed information hours after they first mentally logged it. We’ve discussed the threat of deep recall in depth before, but the short version is this: quality of care decreases, documentation quality decreases, and risk of malpractice increases — a bad combination for healthcare professionals. 

The Conclusion on Speech Recognition Medical Transcription

At the end of the day, speech recognition technology leveraged for medical transcription falls a bit flat. Rather than augmenting a significant portion of the clinical documentation load that plagues healthcare professionals, it simply replaces the task of typing medical notes with dictating them instead. And requires extremely detailed dictation at that. But it’s not all bad, speech recognition technology, combined with AI and natural language processing helps set the stage for more robust medical documentation solutions.

What Clinicians Need to Reduce Their Medical Documentation Load

What clinicians really need is a robust tool that doesn’t require as much heavy lifting on their behalf. A tool that relies on intelligent AI subsystems like machine learning and natural language processing to write and complete their clinical documentation without a draining level of oversight or engagement. That solution is DeepScribe. The first all-encompassing medical documentation solution that can write your medical notes for you.

DeepScribe’s proprietary artificial intelligence can extract all of the medically relevant information from a patient encounter without the highly specific dictation or vocal prompting. Simply activate the application right from your phone, talk to your patient as you normally would, and our system captures the necessary information, categorizes it into the specific SOAP note fields, produces a completed medical document, and classifies and uploads that information into the appropriate fields of your EHR. 

No typing. No extensive dictating. No draining post-session documentation. 

Learn more about how DeepScribe can help you automate your clinical documentation.