Cape Talk, Wednesday night

At rest during the night. There are white papers on how these companies do it. You can read what their methodology is. But in any case, this is the best way to do it. And one of the things they're measuring as well, when they look at you in these sleep labs, is what your heart is doing that whole time. Now, it's not the most relevant data. And so, what these researchers here did, they're from the Department of Biomedical Informatics at the School of Medicine, Emory University in Atlanta, Georgia, in the United States.

Together with some funding from the U.S. National Institute of Health. They looked at this and said, that's a lot of data that's just sitting there not being used well. Can we correlate that to how these people end up doing, how they end up faring in the next 10 years? And they can. What they did is they used deep learning. They took a data set that was just sitting there. So they had, what, 10 years of data polysomnographs to look at? They had, yeah, they had 16,000 pieces of data, people in

The past who had been measured in that way. And they used deep learning to create associations or correlations between what those chest monitors were showing as heart activity and how those people fared over the next 10 years. And what they have established, and they're now beginning with this experiment, but then probably in the future as well, is to assess whether you can look at that data and say this person is more likely to suffer this kind of event in the future. And just some numbers here for you: atrial fibrillation. This is heart.

Heart flutter, actually, I got measured for this one time as well. There can be a 100 to 170% increase in the likelihood that you are going to get that based on the data that is read as part of that sleep lab. They can correlate that. Myocardial infarction, 10 to 40%, strokes, 20 to 40%, heart failure, 70 to 130%, and overall mortality or death, 65 to 80%. I'll quote the end of the paper here. They say deep learning applied to nocturnal ECGs.

With sleep stage information captures cardiac abnormalities, offering a tool for cardiovascular risk stratification. And they go on to say: if you integrate that, you could enhance detection and support personalized clinical decision making. A lot of words to say basically: if you're getting tested for a sleep issue in the future with the help of AI, there's a higher chance that you'll come out of it and find out what your risk profile looks like for a cardiac event. That's just great. That's data that was there.

Software, you may have a doctor talking to you and saying, Hey, look, you are in this category. Let's start dealing with this now, as opposed to what was happening in the past. The data was sitting there and you had no idea necessarily. For right now, could someone say, I'm gonna go and grab a polysomnograph, get that data? Could that person then bring it to a cardiac specialist and they, based on this data, look at it and say, Yeah, maybe we should make some changes here, do some things almost prophylactically?