This week, Amazon Web Services launched a new generative AI tool called HealthScribe, which will allow doctors to use speech recognition machine learning and AI to empower their offices. The goal of which is to help these offices to summarize visits and manage their in-house files.
On the Amazon Web Services page for HealthScribe, it’s said this new tool would help phase out time-consuming documentation. This process done by doctors and staff takes many hours throughout the week. The way HealthScribe will work is that the process would be replaced by an interface that can transcribe, and extract medical terms and medications on the spot.
It will also create summaries from doctor-patient discussions, and then store these newly created health records. To be more specific this tool, will be able to take the conversation between the doctor and patient, and then using AI/machine learning, break it down into quick-study materials.
This would also include the purpose of the patient’s visit to the office, their medical/illness history, and other key pieces of information. All of which would reference the original and full transcript of the visit with footnotes as well.
To protect patient data and manage medical compliance requirements, Amazon says that patient-specific inputs and outputs will not be used to train HealthScribe. They also included that to prove the best effectiveness for the AI, each office would have to train its own language models.
In theory, this could see a growth of LLMs, trained for specific specialties such as internal medicine, ophthalmology, urology, neurology, pediatrics, and others. Allowing for AI models that can provide assistance to doctors with special cases and allow, in the long wrong, the capture of illnesses earlier.
But Amazon Web Services isn’t the first to come up with a program such as HealthScribe. Back in March, Microsoft with Nuance Communications, announced the development of the first AI-automated clinical documentation application. It’s quite similar to HealthScribe.
Either way, Amazon already has several use cases that show a reduction in documentation time, a boost in scribe workflow efficiency, and overall patient consultation recaps.