ISC 2026 | Using AI to improve patient communication and education after stroke

I think that this is maybe one of the best areas that could be used right now. If you just pulled up a standard Claude or GPT, it’s incredible how good they appear to be. They can answer questions. They can give you next steps. They can pull up data. It sounds really convincing. That’s part of the worry because, of course, it isn’t all completely right. But if we use it as a way to question the provider, so in other words, just make us better informed, I think that’s the best way. When a drug is prescribed by your provider, are you sure this is right for me? I have the following other medications or allergies or other conditions. Is this the right thing to do? Asking a large language model what other therapies are being tried in trials in the rest of the world, then you can actually bring up what about trials of this or trials of that? What about new strategies for rehabilitation? For atrial fibrillation, for instance, which is a very common arrhythmia, the most common irregular beating of the heart due to very rapid activity in the top chamber, causing irregular beats in the bottom chamber. There are lots of tools to now, AI tools that are available to better predict if you may be at risk of that, to help identify what medications you should consider using if you have atrial fibrillation, for example, blood thinners, and then what sort of, how you should then search for further help, what type of physician you should go and see, your GP, your cardiologist, your neurologist. So I think that’s all very useful. And again, I think all these things should be taken as an assistant and not as the gospel.

This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.