Trustwise, an AI performance and risk management innovator, has partnered with Health Innovation Kent Surrey Sussex (Health Innovation KSS), Hitachi Digital Services, and Further to co-develop MedAssist GPT, a generative AI application designed to enhance medical education. The tool, leveraging NVIDIA NIM microservices, provides UK medical students with real-time, personalized, evidence-based clinical guidance. Integrating over 700 National Institute for Health and Care Excellence (NICE) guidelines and NHS Formulary data, MedAssist GPT covers more than 30,000 medicines, offering an extensive resource for clinical learning.
Hosted on the Google Cloud Platform, MedAssist GPT utilizes the NVIDIA NIM microservice for the Writer Palmyra-Med large language model (LLM). The application enables students to engage in clinical scenario simulations, practice safe prescribing, and receive instant feedback. Supporting over a dozen languages, it delivers a tailored learning experience while aligning with national healthcare standards.
Advancing AI Safety, Personalization, and Sustainability
MedAssist GPT addresses key AI challenges to ensure high safety and reliability standards. Developers tackled issues such as AI output alignment, data leakage, and hallucinations, ensuring accurate, evidence-based medical guidance that follows regional care protocols. The application’s speed and performance were optimized to provide real-time guidance during clinical simulations, with low latency and high throughput for seamless operation and fast feedback. The platform was designed using NVIDIA AI to balance performance with cost-efficiency, reducing token inefficiencies and promoting operational sustainability by minimizing its carbon footprint.
Collaboration Drives Innovation
The success of MedAssist GPT is attributed to the close collaboration between Trustwise, Health Innovation KSS, Hitachi Digital Services, and Further. Trustwise contributed its Optimize:ai software, enhancing safety, alignment, and responsiveness. Using NVIDIA NIM, Trustwise ensured high-speed inferencing and cross-platform scalability, while Writer’s Palmyra-Med LLM addressed the specific linguistic needs of medical training. Further’s expertise refined the AI models and user experience, ensuring practical utility for healthcare education.
Promising Results from Beta Testing
Early beta testing of MedAssist GPT showed impressive results, including a 90% improvement in AI output alignment with NICE guidelines, a 30% reduction in response times, and a 40% decrease in AI token costs. Additionally, the application demonstrated a 64% reduction in energy consumption and carbon emissions, supporting NHS sustainability objectives.
Moving forward, Health Innovation KSS will continue to collaborate with medical students at the University of Surrey’s School of Medicine to further test and refine MedAssist GPT over the next year. This process will ensure continuous improvement and prepare the application for a full-scale rollout, aiming to enhance medical education and patient care outcomes.
Supporting Quotes
Manoj Saxena, CEO of Trustwise, said, “MedAssist GPT sets a new standard in healthcare training, where safety, speed, and sustainability converge. We’re proud to have created a tool that empowers students with real-time, evidence-based insights.”
Professor Hatim Abdulhussein, CEO of Health Innovation KSS, added, “MedAssist GPT represents a significant leap in medical education, delivering real-time, evidence-based guidance while aligning with our commitment to sustainability and advancing healthcare innovation.”
Paul Watson, VP of Global Healthcare at Hitachi Digital Services, said, “MedAssist GPT is the result of combining our expertise to create an AI application that enhances medical training while ensuring safety and accuracy.”
Mike Gustafson, President of Further, commented, “Through MedAssist GPT, we’ve demonstrated how AI can bridge the gap between vast medical knowledge and clinical decision-making, improving healthcare outcomes.”
About Trustwise
Trustwise helps organizations innovate confidently with generative AI. Its flagship product, Trustwise Optimize:ai, is an AI performance and risk management API that ensures safety, cost efficiency, and operational optimization. Trusted by enterprises in highly regulated industries, Optimize:ai works with any AI model and supports large-scale AI operations. Founded in 2022, Trustwise is headquartered in Austin, Texas, with research labs in Cambridge, UK, and New York.