Translating Medical AI Models into Clinical Practice: Moving Beyond Research

This episode of The Venture Circles highlights a collaboration between Invext and the Siriraj Informatics and Data Innovation Center, focusing on bridging the gap between cutting-edge AI research in healthcare and its practical implementation.

This event covered the following topics:

1. AI Integration for Improved Patient Care
   Approaches to utilizing AI for better diagnostics and treatment.

2. Enhancing Healthcare Delivery through AI
   Effective strategies for using AI to optimize workflows, minimize administrative tasks, and improve operational efficiency.

3. Achieving Measurable Clinical Outcomes
   Techniques for applying AI research to drive tangible improvements in patient outcomes.

Session 1: "Building More Generalization Medical AI"
via Natthawut ‘Max’ Adulyanukosol
Deputy Director of Siriraj Informatics and Data Innovation Center

Thailand is taking its first steps toward integrating medical AI into clinical practice, and data is key. Max showcased how Siriraj Hospital is leveraging its data to prepare for the upcoming wave of medical AI, emphasizing the importance of making these advancements accessible not only at Siriraj but also at smaller hospitals. We invite other hospitals to join a collaborative effort on data standardization by adopting a standard data model.

Session 2: "The MOOVE"
via Dr. Mary-Anne Hartley
Professor, MD, Yale School of Medicine & EPFL School of Computer Science

Massive Online Open Validation and Evaluation for Continuous Real-World Alignment in Clinical Chatbots." We explored how LLMs can democratize medical knowledge despite challenges like licensing, privacy issues, and limited applicability. Meditron-70B, a leading open-source medical chatbot, tackles these challenges with carefully curated guidelines. The MOOVE platform further assesses real-world clinical utility by crowdsourcing expert evaluations—offering participants localized chatbots in exchange for rigorous validation of safety, bias, and trust. By including local clinicians in the evaluation loop, it will ensure the AI model's adaptability to local settings.

Session 3: "From AI Bias to AI By Us, For All of Us"
via Dr. Leo Anthony Celi
Senior Research Scientist, Massachusetts Institute of Technology

AI's impact is raising critical concerns: exacerbating climate change and reinforcing biases to create monopolies. Yet, despite these challenges, AI has the potential to revolutionize healthcare and foster systemic, integrated thinking. Realizing this potential means more than just developing new models; it requires a fundamental redesign of our knowledge, education, and innovation systems. Traditional solutions may not suffice in the age of AI. We need every stakeholder involved—doctors, nurses, developers, students, and even patients—in planning AI solutions. By collaborating together, we can create AI that truly works for all of us.

In addition to insightful speeches from our speakers, we also had a lively conversation about medical AI, focusing on pushing medical AI into real-world clinical practice. The audience actively joined the discussion, sharing their unique perspectives and experiences. This event marks a small, but significant step toward Thailand's journey in applying medical AI to improve patient health. We are glad to be a part of this step, and we look forward to a future of continued progress and innovation.

Author: Sornchai Boonwatcharapai, SiData+ Blog Team

Editor: Sasinipa Uthaisaad, SiData+ Blog Admin