The world’s leading publication for data science, AI, and ML professionals.

Medical Transcription in the Age of Voice-Tech

Artificial neural networks are slowly revolutionizing the medical field through automatic speech recognition technology for medical…

Physicians are now readier than ever to focus on treating patients than devoting time to endless administrative tasks – why? Because of AI. Artificial neural machine networks (ANN) are slowly changing industries not just in data science but in both the medical field and the corresponding translation industry. Today we’ll take a look at the use of AI and translation technologies in medical transcription, to see how data science, medicine, translation are all working together for the general public. We’ll take a look at ANN’s impact on the medical field, particularly automatic speech recognition (ASR), and the language barriers of ASR in patient-doctor interactions.

Image by Freepik.com
Image by Freepik.com

Artificial Neural Networks in the Medical Field

Artificial neural networks work sort of like the neurons in the human brain, which is what they’re modeled after – neural machine networks have the same interconnectivity and are capable of new "neurons" of the brain, but the similarities end there.

We’re learning that, since machine intelligence programs learn very quickly when loaded with lots of data, though, they can now parse much contextual information from that. They may not be able to know contextual cultural information, but they respond well to contextual visual clues, according to medical research. As this article poses, software algorithms are learning to overtake human intelligence, and they’re becoming good at learning patterns through layers of incrementally increased huge data over the long term.

With the pandemic of 2019, the significance of AI to solve problems within the medical field, and then relay this information to the general public by virtue of translation, took on an urgent note. Artificial neural networks (ANN), as building blocks of AI tools for virtual physicians, medical transcription, and diagnosis identification, is now not only a subject of medical research. ANN is now implemented widely through the use of automatic speech recognition (ASR) for medical transcription in the international field. That’s the effect of medicine, tech, and translation working together for the public good is still being felt today.

Medical Transcription as Medical Scribes with ANN

That said, artificial neural networks (ANN) are revolutionizing medical research if not the medical field. They may very well be used for digital imaging, identifying new diseases, and more. Rather than using AI to create a virtual physician with no need for human doctors, the most popular field that’s emerging from ANN research is one that calls for a successful collaboration between humans and machines: medical transcription. Medical transcription is a specialized medical service that’s slowly adapting to the changes in the medical profession.

Automatic speech recognition (ASR) technology, built with ANN, (used often in the translation industry) is rolling out to the medical field, for use by doctors who want to dictate directly to both nurses and patients. These are virtual medical scribes, as opposed to medical transcriptions which humans. By recording the transcription of the doctor’s interaction with the patient, speech-recognition technology can automate doctor’s visits, without the tediousness of piles and piles of documentation.

Medical transcription with ASR helps in everything from updating the patient’s Electronic Health Record (EHR) to automating medical charts to marking medical appointments to setting up doctor-to-doctor referrals. Examples include Saykara, owned by Nuance Communications, a Seattle-based startup. They market Saykara as the first "mobile AI assistant to automate physician charting."

Voice-recognition has been around since the 1950s, but like all great technologies, did not find its use until the last few decades, particularly with the debut of Google’s Voice Search and Apple’s Siri. Now there’s Amazon’s Alexa, Microsoft’s Cortana, and many others. Otter.ai, Augnito, and Decrypt are just some of the newer brands in this voice-tech industry, revolutionizing clinical workflows through medical transcription.

All well and good, right? Yes, and no. Automatic speech recognition (ASR) has had a major impact on translation technology. This study proposes two problems: 1. Audio transcription and translation are known to be more time-consuming than text translation and do not actually automate the process. 2. Audio quality can become poor due to background noise, overlapping voices, and other conditions, which translation technology has not provided the best solution to yet. And 3. language barriers, which could be resolved with translation services have specialties in AI and medical documentation.

So, AI-based transcription tools and medical documentation go hand-in-hand, with these tools listening in on the patient-doctor interaction without any need for hand-written documentation. But what happens when, say the patient is speaking in Korean and the doctor is speaking in English, as this study points out. In Korea, speech recognition technology is more difficult, when there are language barriers to solve, according to the study.

In addition, there are varying degrees of language problems to solve with ASR: from code-switching from one language to another to speaking in unknown dialects, to being limited by few language pairs, to recognizing cultural contexts. That’s why it’s best to utilize the services of a medical transcriber from the translation industry, since those in that industry have specialized linguistic expertise, technical medical knowledge, and are fully equipped with AI tech.

So, the verdict is out: in ASR, AI is not equipped to fully transcribe a multilingual patient-doctor interaction, yet. ANN might be changing that, soon, but not just yet. Hopefully, speech recognition gets more advanced as this technology becomes more commonplace in hospitals and clinics, in an international capacity.


Written By

Topics:

Related Articles