Automated Medical Transcription

Precision Agriculture
March 10, 2020
Self-Sovereign Identity
March 10, 2020

Automated Medical Transcription

Dictating patient notes is a core task in a clinical practice. Artificial intelligence promises faster transcriptions, as well as real-time diagnostic analysis.

Key Insight

Dictating patient notes is a core task in a clinical practice. Artificial intelligence promises faster transcriptions, as well as real-time diagnostic analysis.

Why It Matters

Transcribing recordings is a tedious process that relies on excellent sound quality and a good understanding of medical terminology—not to mention confidentiality. Making the process more efficient and effortless could improve operations throughout hospitals and medical practices.

Examples

In order to meet compliance regulations, a strict protocol must be followed for the transcription to be legally performed by a third party. But what if the transcription is performed in real-time? Not only would it be easier and more cost-effective for building patient records, but an additional layer of machine learning could help doctors learn even more about their patients during visits.

What’s Next

In December 2019, Amazon launched Amazon Transcribe Medical and a companion service called Amazon Comprehend Medical. Both are intended to make medical practices more efficient. Transcribe Medical does what its name says: It runs voice recognition on audio of doctor-patient interactions and transcribes the conversations directly into an electronic medical record. In a private practice without medical residents, this frees the doctor from having to move between the patient and her computer to enter symptoms and other information. Comprehend Medical is intended for developers to help them use unstructured medical text in diagnostic tools. AWS’s software is designed to be integrated into devices and apps using an API. Microsoft’s Azure and Google Cloud are also working on similar systems.

The Impact

The biggest challenge for tech companies will be proving the compliance and accuracy of their systems. Imagine an error in which “hyperglycemic” (high blood sugar) is mistakenly recorded as “hypoglycemic” (low blood sugar).

Watchlist

Alphabet, Amazon, Apple, Google, IBM, Microsoft, Nuance, Stanford University, healthcare providers, hospitals, and government agencies.