I think one of the things when we’re talking about digital biomarkers, we need to be actually very careful about are we talking about a digital biomarker or are we talking about a digital phenotype? So we do a lot of digital phenotyping right now because anything that’s collected through a digital device could be called a digital phenotype. Now the issue is, how do you convert that digital phenotype into a digital biomarker? And I think that the idea of a digital biomarker is something that still has to be validated, so just in the way that we have imaging biomarkers, we have now blood-based biomarkers, you know, they’ve all gone down this validation process...
I think one of the things when we’re talking about digital biomarkers, we need to be actually very careful about are we talking about a digital biomarker or are we talking about a digital phenotype? So we do a lot of digital phenotyping right now because anything that’s collected through a digital device could be called a digital phenotype. Now the issue is, how do you convert that digital phenotype into a digital biomarker? And I think that the idea of a digital biomarker is something that still has to be validated, so just in the way that we have imaging biomarkers, we have now blood-based biomarkers, you know, they’ve all gone down this validation process. And in many cases, they’re FDA approved or FDA at least certified. And I think that to get to digital biomarkers, we’re going to need to go through that same process to get from what’s really a digital measure to a digital biomarker. So that’s first of all. But then how will these digital biomarkers be different from a traditional biomarker? So I think that the difference will be that these digital biomarkers, in my mind at least, are going to be sort of multi-sensor. I don’t think they’re going to be measured by just one device or one measure. I think in fact what they’ll do is there’ll be a triangulation of multiple different digital data types that come together into a measure that’s going to get to that digital biomarker validation process because you need to be highly accurate, right, in terms of what’s that context of use in which you are developing that digital biomarker for. I suspect that these digital biomarkers are also not going to be static. So, for instance, if we take a blood-based biomarker, it’s usually there’s some threshold in which you say, oh, yes, there’s disease, right? This would be an AD biomarker positive versus an AD biomarker negative. So there’s some threshold measure. I don’t think that’s going to be the case with the digital biomarker. I think what’s going to happen is it’s going to be very much more of a dynamic signal. And it’s going to be so dynamic that even the sensor information that comprises it is shifting and changing. But nonetheless, I think that that signal, even though it’s vacillating and moving, I think it will end up being an accurate representation of what you’re trying to measure at a level that fits the context of use that we typically think of as a biomarker indicator.
This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.
