AWS Brings Generative AI to Medical Transcription Apps
Now in preview, HealthScribe will let software developers build HIPAA-compliant transcription applications for medical settings.
Amazon Web Services has just debuted a new generative AI-powered transcription solution aimed specifically at healthcare software developers.
Now in preview, AWS HealthScribe is an API-based service that will help software developers use AI to build transcription tools that meet the compliance, security and privacy requirements specific to the medical industry.
The service was described by AWS as a "HIPAA-eligible service that empowers healthcare software providers to build clinical applications that use speech recognition and generative AI to save clinicians time by generating clinical documentation."
The underlying technology in AWS HealthScribe is Bedrock, which AWS launched this past spring. Amazon Bedrock lets developers build generative AI projects on the AWS cloud that leverage "foundation models" (AWS' version of LLMs, or large language models, the likes of which power OpenAI's ChatGPT).
HealthScribe promises to enable software to turn spoken conversations between, for example, a doctor and their patient into properly sourced medical notes and other documentation, which can then be entered into the provider's electronic health record (EHR) system.
Normally, this is a time-consuming process for providers, and while text-to-speech technologies already exist to alleviate some of the complexity, a generative AI-powered solution, argues AWS, is ultimately needed to turn transcripts into data that's easily digested by EHRs.
Of course, building, training and securing an LLM on a medical provider's specific fields -- accounting for differences in jargon, medicines, dosages and compliance requirements -- is just as, or even more, time-consuming.
According to AWS, HealthScribe will to do this heavy-lifting for healthcare developers:
By integrating AWS HealthScribe into a clinical application, healthcare providers can leverage built-in speech-to-text capabilities to create robust conversation transcripts that identify speaker roles and segment transcripts into categories (e.g., small talk, subjective comments, or objective comments) based on clinical relevance. The application can then use AWS HealthScribe's NLP and generative AI capabilities to extract structured medical terms, such as medical conditions and medications, and generate discussion-based notes that include relevant details (e.g., key takeaways, reason for visit, and history of the present illness) that a clinician can review and finalize in their EHR.
The current version of HealthScribe is trained for general medicine and orthopedics, with more specialties to come.
"Documentation is a particularly time-consuming effort for healthcare professionals, which is why we are excited to leverage the power of generative AI in AWS HealthScribe and reduce that burden," said chief of machine learning and AI at AWS Bratin Saha. "Today's announcement builds on AWS's commitment to the healthcare and life sciences industry and our responsible approach to technologies like generative AI to help reduce the burden of clinical documentation and improve the consultation experience."
More information on the HealthScribe preview is available here.