With time saving aspect etched deep, medical transcription that is automated is an express facility to doctors giving sharp tech sharks a thrust for innovation,
Conventionally to doctors, note-taking to fetch EMRs (electronic medical records) has always been a burdensome experience and clerks with perfected medical knowledge have been sought for the purpose. But with advancement in technology, such a scenario undergoes a change as scores of hospital systems have now seen embedded with AI technology.
Nonetheless, doctors don’t feel entirely comfortable with such tools and complain of inaccuracy and an ER doctor Bon Ku serving at Thomas Jefferson Hospital University and who also looks after the Health Design Lab of this university as Director confesses, “If there were a really smart voice transcription service that was 99% accurate, I would definitely use it.”
Major players in tech industry have been working hard for the past several years, for that distinct glory of providing the maiden tools that doctors are desperate for.
Recently, that is in November this past year, Google founded a software which is an open-source machine learning based and as a result of which, doctors are empowered to gather valuable analysis out of medical records. There are, in fact, planted a couple of programs, i.e. API meant for natural language processing linked to healthcare and which is tasked with:
- Analysis of critical info present in patients’ medical history,
- Arranges such key facts in proper format, and
- Presenting a summary for quick actions by doctors.
Also, key info can be garnered from scores of sources, such as medical records and / or notes which are transcribed. The ultimate objective is to craft a swift way ahead for healthcare professionals to carry out a quick review upon patients’ medication and complication of the past.
The second in the series is AutoML Entity Extraction again meant for healthcare, which is actually a low-code tool kit assisting doctors to fetch absolute data from voluminous health records. Now, we have every reason to feel jubilant as both these tools are made available for FREE for healthcare professionals, insurers and biomedical companies but only till December 10, 2020.
Not mincing words, but major tech companies simply tend to rustle out an amenable and palatable platform for doctors to quickly record their patient-conversation without requiring a computer to type into and great amount of efforts and sheer interest in medicine technology centre around this. Towards the objective, success (to a certain degree) has been achieved by Amazon, Microsoft and Google as software are successfully developed while tools are also being designed for healthcare framework with the notion of revenue generation.
Latest in the series, is Nvidia, which otherwise remained transfixed on imaging technology but now drifted into medical transcription. In early months of past year, a service offering was made by Nvidia which was called BioMegatron which is meant to identify speech discourse. Such data set is said to have been sampled on 6 billion medical terms and comes at 92% accuracy. Besides, there are still dozens of companies such as Dragon, MModal, Sukai Ai and Saykara which are into medical transcription enabling efficiency of doctors.
Transcription, triggered by AI, is not lesser than a milestone in the otherwise lengthy journey of automated medical processes resulting in doctor’s work getting electronic and scores of healthcare professionals now rely upon computer systems to pull up the patient records. Regarding ER duration, going by the 2013 paper study, it suggests that doctors make over 4000 clicks during their shift. But there are healthcare professionals who relied upon EPIC electronic health record system that also have a “dot phrases” program which speed up the note making and retrieving patients’ info (EPIC do come with an AI transcription module). Such a dot phrase facility does give way to disease generalization, as frequently typed entries are enabled pertaining to symptoms and other health complaints, which is otherwise perfect from medical billing point of view. As fallout, it is seen that doctors fail to fathom the patients’ last visit objective.
Ku complains, “Most of patient records are garbage—they’re full of templates. “Ninety percent of our diagnoses come from the interview; it doesn’t come from diagnostic imaging or lab tests. It’s about me being able to get the story from my patient—but that becomes hindered because there’s this insane pressure to enter data into a computer.”
Finally, to update EHRs, doctors have to invest time and efforts in putting data into and Ku playfully regards it as “pajama time” indicating time-length wasted in recalling patient records in system. As such, doctors are seen rallying in favour of digital note-taking phenomenon which is simply like talking to Alexa. Arguably, Ku observes that any such framework which would filter data of patients from a discourse or any facility to prescribe medical test would be the desired extension as doctors’ time and mingling with patients will extend tremendously. Having said that, technology still needs to be driven to the point where there is a perfect harmony with doctors, i.e. accuracy as doctors need not spend time rectifying what AI got incorrect.
Ku expects, still, “There has to be some safety mechanism”.