Amazon has been investing heavily in both healthcare as well as generative AI technologies for an extended period now. Amazon’s AWS has recently brought the two niches together by launching HealthScribe—a tool designed to assist clinicians and medical practitioners with documentation and note-taking. Medical professionals spend a considerable amount of time documenting case details, patient information, procedural nitty-gritty, as well as legal information. Tools like HealthScribe bring a potent AI solution to medical writing and documentation, cutting out the time spent on these tasks and allowing healthcare workers to focus better on their primary responsibilities—patients. Amazon’s efforts also come at a time when the firm, like other competitors, has been developing language models. Bedrock, Amazon’s own LLM witnessed release through AWS’ platform, which features numerous foundational models including Amazon’s own.
Apart from Amazon’s offerings in the healthcare space, Google, too, has been looking at integrating AI with healthcare applications and has recently been partnering with numerous hospitals and independent practitioners to test its Med-PaLM 2 language model tailored to assist doctors and healthcare professionals. Evidently, healthcare AI is witnessing considerable growth. Tools like AWS HealthScribe also bring about a new approach to supplanting mundane tasks with a cost-effective and time-saving solution by helping their human counterparts focus on tasks that require critical thinking and problem-solving skills. The upcoming sections explore Amazon HealthScribe and its features.
Amazon’s Generative AI Push and AWS HealthScribe
HealthScribe is currently being used primarily by healthcare tech firms and pharmaceutical companies.
Amazon has recently been enhancing its capabilities in natural language processing, LLMs, and generative AI architectures. AWS HealthScribe is positioned as a tool to aid the long and arduous process of updating electronic health records manually. HealthScribe can also be utilized as a base model to build applications that support functions like recording patient information, clinical visit summaries, case details, and pharmaceutical data like prescriptions. Along with attributes like AI writing, HealthScribe also features intuitive capabilities like speech recognition and speech-to-text conversion protocols, which can make clinical documentation rather straightforward. HealthScribe’s AI capabilities can also allow doctors to detect key details within summaries and can seamlessly transfer them to electronic health records. Currently, HealthScribe remains limited to two medical specialties—General Medicine and Orthopedics. AWS might extend support to other specialties if it receives favorable reviews from medical professionals who work in these niches of healthcare. The success of the preview in these two spaces will also determine the overall extent to which HealthScribe can be primed and offered to other departments.
More importantly, HealthScribe is also in line with the Health Insurance Portability and Accountability Act (HIPAA). This makes the novel AI tool HIPAA eligible and ensures patient information and data will remain protected. Presently, Amazon HealthScribe’s medical AI support and features come for a subscription price. The model is unique, and the users are billed based on the number of seconds they have used up in transcription time. The price per second of transcription is pegged at $0.001667 and a minimum per-request charge of 15 seconds. The pricing also includes other features like speaker-role identification, turn-by-turn transcription, clinical summarization, transcript segmentation, and clinical data extraction. Features like this might also enhance the role of analytics and big data in medicine. As HealthScribe garners popularity, more medical facilities are expected to enter the foray with their requirements and suggestions, potentially making the application more robust and in line with responsible AI requirements.
Amazon HealthScribe’s Scope and Technical Features of the Generative AI Tool
Healthcare AI apps are gaining considerable traction since the AI boom
AWS HealthScribe presents numerous opportunities for clinicians and healthcare firms to optimize their workflows with efficient record-keeping. Currently, Amazon has not approached doctors and hospitals directly with its tool and has instead partnered with healthcare technology and pharmaceutical firms to assess the tool’s compatibility and functionality within dedicated departments in these firms. As general physicians and orthopedic surgeons from partner firms comment on the utility of the application, AWS is expected to make suggestions and necessary changes to better suit the healthcare framework. The transcription and AI-supported service is capable of helping clinicians sift through aspects like the chief complaint, history of presenting illness, clinical symptoms, family and medical history, and more. AWS HealthScribe is also available through API for potential customers looking to plug it into the framework for other healthcare-related necessities. Similar AI-supported transcription features are also offered by firms like Otter.ai, albeit for a non-medical setting.
The application essentially transcribes the conversation and then presents structured information through AI-generated content. AWS HealthScribe quotes the source of its information in the original transcript upon presenting aspects like clinical history and other segmented parts of the doctor-patient conversation. Major clients that plan on adopting Amazon’s medical AI service include giants like 3M and Babylon, who have been eager to optimize their workflows based on the latest technologies available. Medical transcription firms like ScribeEMR have also been quoted as being interested, given that their firm comprises numerous partner physicians who use their services for record-keeping.
The Outlook for Generative AI in Medicine
HealthScribe will be extended to more departments in due time.
Progress in AI is bound to result in enhanced tools across various disciplines. This also includes sensitive and high-stakes professions such as medicine and healthcare. Amazon HealthScribe offers a potent alternative to conventional transcription options and services, by using AI protocols to extract information directly from transcripts and entering the same into easily transferable files and documents that can be added to patient EHRs. These tools will invariably also impact important sub-fields of medicine such as population health and community healthcare. As AI tools expand their reach in the healthcare space, AI safety and key regulations surrounding the use and deployment of artificial intelligence will have to be ratified to ensure all-around data privacy and protection.
FAQs
1. Is Amazon HealthScribe free?
Amazon HealthScribe is a paid service that charges users based on the seconds of transcription consumed. Each second of transcription is priced at $0.001667, with a minimum transcription rate of 15 seconds.
2. How does AWS HealthScribe provide information about patients?
Amazon’s HealthScribe framework uses generative artificial intelligence to pull from transcribed conversations to offer pointed information surrounding key aspects of medical conversations such as clinical symptoms, chief complaint, and medical history among others.
3. Is HealthScribe used in hospitals?
Amazon’s AWS Healthscribe is currently being offered to large healthcare technology and pharmaceutical firms to test the framework. Currently, the services are being primed in the General Medicine and Orthopedic divisions of these firms and their partner establishments.