Deepgram has announced NOVA-3 Medical, an AI-to-text (STT) model tailored to transcription in a healthcare demanding environment.
NOVA-3 Medical is designed to seamlessly integrate with existing clinical workflows and aims to address the growing need for accurate and efficient transcription in the UK’s public NHS and private healthcare environments.
As electronic health records (EHRS), telehealth and digital health platforms become more popular, the demand for reliable, AI-driven transcription is more than ever before. However, traditional speech to text models often struggle with complex, specialized vocabulary used in clinical settings, leading to errors and “haptic illusions” that can compromise patient care.
Deepgram’s NOVA-3 Medical is designed to overcome these challenges. The model leverages advanced machine learning and specialized medical vocabulary training to accurately capture medical terms, acronyms and clinical terms, even in challenging audio conditions. This is especially important in environments where medical professionals are away from recording devices.
“NOVA-3 Medical represents an important leap in our commitment to transforming clinical documentation through AI,” said Scott Stephenson, CEO of Deepgram. “By addressing the nuances of clinical language and providing unprecedented customization, developers can now build products that improve patient care and operational efficiency.”
One of the key features of the model is to provide structured transcription that seamlessly integrates with clinical workflows and EHR systems, ensuring critical patient data is accurately organized and easily accessible. The model also offers flexible, self-service customization, including key theme prompts for up to 100 key terms, allowing developers to tailor solutions to the unique needs of various healthcare specialties.
Versatile deployment options including on-premises and virtual private cloud (VPC) configurations – ensuring enterprise-grade security and HIPAA compliance. This is important to meeting UK data protection regulations.
“The speech-to-text for enterprise use cases is not trivial, and there is a fundamental difference between voice AI platforms designed for entertainment and entertainment use cases.” “Deepgram’s NOVA-3 and NOVA-3-Medical models are the leading voice AI products, including TTS, in terms of the accuracy, latency, efficiency and scalability required for enterprise use cases.”
NOVA-3 Medical Benchmark: Accuracy, Speed, Efficiency
Deepgram has conducted benchmarks to demonstrate the performance of the NOVA-3 Medical. The model claims to provide industry-leading transcriptional accuracy and optimize both overall word recognition and critical medical terminology accuracy.
Word Error Rate (WER): With a median WER of 3.45%, NOVA-3 healthcare is superior to its competitors, with a 63.6% reduction in error compared to its next best competitor. This enhanced accuracy streamlines your workflow by minimizing manual modifications. Keyword Error Rate (KER): Importantly, NOVA-3 Medical achieves a KER of 6.79%, reducing errors by 40.35% compared to its next best competitor. This ensures that key medical terms such as drug names and conditions are accurately transcribed, reducing the risk of misunderstandings and patient safety issues.
In addition to accuracy, NOVA-3 medical care is superior for real-time applications. This model is ideal for telehealth and digital health platforms as it transcribes speeches 5-40 times faster than many alternative speech recognition vendors. Its scalable architecture ensures high performance even with increased transcription volume.
Furthermore, NOVA-3 medical care is designed to be cost-effective. Starts at $0.0077 per minute for streaming audio. This argues Deepgram’s claim is more than twice as affordable as a major cloud provider. This allows healthcare tech companies to reinvest in innovation and accelerate product development.
Deepgram’s NOVA-3 Medical aims to enable developers to build transformative medical transcription applications, driving exceptional outcomes across healthcare.
(Photo: Alexander Singh)
Reference: Autoscience Carl: First AI Scientist Writing a Peer-Reviewed Paper
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expo in Amsterdam, California and London. The comprehensive event will be held in collaboration with other major events, including the Intelligent Automation Conference, Blockx, Digital Transformation Week, and Cyber Security & Cloud Expo.
Check out other upcoming Enterprise Technology events and webinars with TechForge here.
