In this Q&A, Mohammed Saeed, MD, PhD, chief medical officer at Solera Health, explores how wearable devices and continuous remote monitoring are reshaping clinical oversight, from early intervention to AI-driven pattern detection.
Watch parts of the full video interview with Mohammed Saeed, MD, PhD, on the Applied Clinical Trials YouTube channel, where he goes deeper on wearable integration, remote monitoring workflows, and the role of AI in clinical oversight. Subscribe for more expert conversations shaping the future of clinical operations and drug development.
As wearable devices move beyond fitness tracking into continuous clinical monitoring, the challenge for healthcare providers is no longer access to data but managing its volume, velocity, and reliability without adding to an already overwhelming workflow.
To explore this further, Applied Clinical Trials spoke with Mohammed Saeed, MD, PhD, chief medical officer at Solera Health, about how remote monitoring is reshaping care delivery, what stands in the way of meaningful clinical integration, and where artificial intelligence (AI) fits into the future of patient oversight.
ACT: What is most significant about wearable data becoming part of routine clinical decision-making today?
Saeed: I think that wearable data gives physicians and other clinicians a snapshot into the life of their patients outside of the clinic and hospital, and how their overall health holds up to real world stresses. That's something you can only get from wearables data, and I think that's a very unique aspect that can help you better understand what limitations patients are going through and identify patients that might be having challenges and need earlier interventions.
ACT: How is continuous, remote monitoring changing how care is delivered and managed between visits?
Saeed: Continuous remote monitoring really allows caregivers to have insight into what's going on with patients at home and identify early warnings of deterioration, so that it's possible to intervene before the health condition gets to the point where they need to be hospitalized or go to the emergency room.
For example, if doctors or nurses see some variables going in the wrong direction, they can simply give a phone call to the patient at home, make some adjustments to medications, and see if the response goes in the right direction. If it does, that's a great course correction. If those interventions don't help, then they might have to escalate further. But early intervention, I think, is the key emphasis with remote monitoring.
ACT: What challenges remain in integrating wearable data into clinical workflows in a meaningful way?
Saeed: One of the challenges with data from remote monitors and wearables in general is the data overload, or information overload, problem that providers are dealing with today. When you look at the world of EMRs and the amount of data we have to deal with on a day to day basis, it is overwhelming. And when you add to that data coming from wearables — streamed at very high velocity, continuous data that can tell you things about heart rate, number of steps, sleep patterns, glucose levels, activity — and you have to combine that with everything else while taking care of large cohorts of patients, that's a big challenge.
And then on top of that, remote monitoring can generate alarms and alerts that are frequently false, and that's where you get a kind of cry wolf syndrome. We have this problem in hospitals too, where alarms are so often false positives that we become numb to them. And when a patient has a life threatening condition, because we're so used to the alarm being false, we don't pay attention to it. Similarly, when home alerts from remote monitoring are frequently false, you can start ignoring potentially dangerous alerts that are being set off.
ACT: How do you see FDA-cleared devices and reimbursement models influencing broader adoption of wearables in healthcare?
Saeed: I think there are two aspects to this. From the FDA perspective, as long as these devices undergo rigorous evaluation, that helps build trust that providers can have in them. It's absolutely important for wearables to be utilized as part of clinical care that there is verifiable trust in those devices. The FDA's role in building that trust between the vendors that make these devices and the providers that would use them in clinical care is absolutely paramount.
The second aspect, from a reimbursement perspective, is building models of care where the time providers spend reviewing that data is factored in. These data streams can be very large volume, and how you factor in the time spent — and how that influences care — really matters. Building reimbursement models that make sense is also very important.
ACT: Looking ahead, how might AI and wearable data together reshape patient care and clinical oversight?
Saeed: AI is making major advances in being able to parse through these massive data streams. Going back to that information overload problem — it's just not feasible for a provider or group of providers to be taking care of hundreds or thousands of patients with all this data being streamed from their homes and identify the subtle trends. By building AI models that can identify trends in data and separate true alarms from false alarms, you can help focus providers' attention on what matters most.
And because these data sets, when you factor in everything else about the patient — their pre-existing chronic conditions, lab data, imaging data — if you can build AI models that factor all of that together, they may be able to identify subtle patterns that providers themselves would not be able to detect. There may be subtle changes in signals that are actually very indicative or predictive of future bad outcomes.
For example, people have done some really interesting work looking at how speaking patterns might change as a predictor of future heart failure exacerbations — things that are so subtle it would be very difficult for a physician to detect, but with very smart AI, you might be able to analyze just a few sentences and identify whether those types of changes are predictive of heart failure. Similar approaches can be applied in behavioral health, depression, anxiety, and many other diseases. Those kinds of subtleties are where AI can become a very interesting tool for discovering new ways of identifying impending deterioration in patients.
Editor's note: This transcript is a lightly edited rendering of the original audio/video content. It may contain errors, informal language, or omissions as spoken in the original recording.
