What technologies will shape the future of healthcare?

Chris_Wynder
Chris_Wynder E Community Moderator
edited November 18 in Articles #1

TL;DR Anything that automates processes. This is why Next Gen AI/LLMs are so attractive.

Back-end workflow/Decision automation. Connecting across the 3 segments of Care management (Primary, Specialty, Hospital) is the best bang for the buck. Technologies that integrate data sources and provide appropriate information for the role in clinical workflows- across the segments are the “low-hanging fruit”.
Industry specific Language Models AND the support data management layer. This is probably IMO the coolest technology that we are seeing academic projects and some very nascent start-ups. The premise is: If instead of training a GPT on the wide corpus of human knowledge and allowing it to synthesize that knowledge as a generalist; What if we went smaller focused “LM” around a “Medical knowledge”?

There is a new AI framework called RAG (Retrieval Augmented Generation). Which basically means a Hospital could have a private Language Model that defines their core knowledge; how they run the hospital, what their patient population looks like, what their historical data on patient outcomes, etc. This private “LM” would then be the first stop for internal users. This would also act as a gatekeeper ensuring that patient information is never sent outside of the organization. Using a RAG framework users could query a large public Medical LM securely combining their corpus of knowledge with the best of medical knowledge worldwide. This has the largest potential for Healthcare positive change. This would give the benefits that the average person sees from ChatGPT but without the security risk or hallucinations.

This is really attractive in Canada as we have as a rule been quick adopters of healthcare tools that align to worldwide best practices.

Cyber-Security AI. The biggest roadblock to innovation in healthcare is data security/privacy risk. There are some really interesting new uses of general AI to identify risks in data architecture and connections and recommend remediation. This is seems boring- but most of healthcare’s problems are highly fixable with existing UX and data management technology. The concern is always “this safe enough for healthcare?”. The idea behind Cyber-security AI is that is you can literally ask that specific question and get an answer AND the fix.

Want to learn more?

OpenText 2023 Patient Experience hackathon (video tutorials) and our Hackathon community are good places to start. Join our 2024 hackathon on Sustainable Medtech

Resources and citations

Flipboard magazines

AI orchestration: 

https://flipboard.com/%40chriswynder/ai-orchestration-h35a5j99z?from=share&utm_source=flipboard&utm_medium=share

Chris Wynder

Chris is a Director of Product Marketing working with our Developer product team and community. He has a wealth of information management knowledge, particularly in highly regulated industries. He shares his deep belief in analysis and taxonomy as the basis of good information governance in his blogs.