Steven Steinhubl, MD, is the Chief Medical Officer for physIQ. His work at physIQ has focused on the implementation of clinical programs built specifically around the novel capabilities of digital technologies, in particular, personalized predictive analytics.
Can you describe your career journey and how you ended up at physIQ?
I started as a chemical engineer at Eastman Kodak, right at the beginning of Kodak’s digital disruption. At Kodak, I worked at a volunteer ambulance where I fell in love with healthcare. After medical school, I did some training in the Air Force, was an internist before doing my cardiology fellowship at Cleveland Clinic, and then went back to the Air Force.
Before I went into medicine, I felt like doctors knew everything. And so I was naively surprised, in medical school, how much remained unknown in health care. During my residency, I found that there was a general tendency to ignore the stuff we didn’t know and focus on the things we knew. Perhaps because of my engineering background, I felt strongly the opposite: I was fascinated by all the things we didn’t know.
When I was at Cleveland Clinic, I met Eric Topol. He was the chairman of Cardiology at the Cleveland Clinic. And in working with Eric, that desire to question and explore the unknown was reinforced. He was the first person I met who was passionate about continuous health care improvement, and that research was a critical component of that. When I was at the University of Kentucky, after I finished my Air Force commitment, it really hit me that we’re doing this amazing work in the hospital, putting in multiple stents and saving lives, but then once somebody goes home, they are often completely forgotten about. As great as the many things that we can do when somebody is sick and in the hospital, the fact that so little of care is directed towards keeping people well inspired me to change the focus of my research into changing systems of care to improve health.
How did you pursue that focus in continuous monitoring?
As it turned out, Eric had gotten funding from Qualcomm Life to start a digital medicine group at Scripps, which he had gone to after the Cleveland Clinic. So I was lucky enough to come there as the Founding Director of the Digital Medicine Department at Scripps Research Translational Institute. In that role, I first got to know the physIQ team. They had a similar vision and were asking the same question, “How do you keep people well?” They approached it in a very evidence-based, data-driven way, which I found to be very rare among digital health start-ups.
Later, we were able to carry out a very innovative Scripps-physIQ collaborative project in an Ebola treatment center in Sierra Leone, funded by the White House, CDC and USAID. In the performance of that very challenging project, physIQ just blew me away with how solution- and data-driven they were. We created state-of-the-art, continuous, wireless, vital sign monitoring capability in the middle of rural Sierra Leone. We were able to completely transform how we kept patients and health care workers well, and more rapidly recognize and manage decompensation in individuals with Ebola.
What was the impact of that work?
Over the last several years, but particularly with COVID, major gaps in current systems of care have been brought to the forefront. COVID was also the first time for many people to see what was possible when you focus on individual changes in health rather than population metrics.
Only by being able to recognize what somebody’s “normal” is and identify the earliest deviation from that could we recognize COVID much sooner, and sometimes before there were symptoms. And so I’m hopeful that, of all the horrible things that have come out of COVID, that one of the positive things will be a recognition of what’s possible through continuous remote patient monitoring.
How adoptable is this for clinical research and what needs to happen to get to that point?
Here is the challenge that we face so far. There are a lot of people who have created these great solutions; physIQ is just one of them. But they’re creating them in this way that just assumes that participants in trials have no problem wearing a monitor, have no problem charging it every day, or downloading something or stopping their day and measuring something several times a day. We create all these solutions, assuming that all people are okay with altering their life in order to monitor themselves. And in some situations, they might be more willing to do that. But the important thing is that there is no one digitally enabled solution that fits all, it’s all based on the use case: I believe that the best sensors are completely passive ones, or at least as close to passive as possible to still acquire the data needed.
One of the biggest hurdles to continuous monitoring working in both clinical trials and clinical care is the fact that once you actually design something that people will use, all of the data ends up being dumped into existing systems of care or existing clinical research infrastructures that are not designed to do real-world, real-time data monitoring. And so what happens is that overly busy clinicians are overwhelmed beyond what they already are, and are unable to take advantage of the advantages that digital technologies provide.
Any solution has to begin with asking, “What’s the problem we’re trying to solve?” After starting with a clear understanding of that, the next questions are: “Do the sensors and wearables exist to do that? Are there ways to successfully implement them? Can we reach the population that we need to be able to reach? Can we adequately communicate the ongoing results in order to provide value to the people who are participating?”
We have to get away from “episodes of care,” as much as possible in both research and care. If you have digital connectivity, when somebody has some information, or they want to ask a question, we have to be able to react to that need in as close to real time as possible. Therefore, you must build that 2-way digital infrastructure. Not only do you need a system that ingests the data, you also need sophisticated data analytics to make sense out of it to enable rapid communication of information back to participants and patients.
What are the digital trends in healthcare you’re seeing that could impact clinical research?
The two trends I see, which are overlapping, are the sophistication of wearable sensor devices, and the growth in the number of people wearing them. If you look at a lot of the work in COVID, especially in the larger studies, participation was based on the ability to “Bring Your Own Device.” While that worked great for rapid, proof-of-concept work, it was also quite exclusionary, as we were only studying individuals who had both the interest and income to buy a smartwatch or fitness band. Clearly, that excludes too large a percentage of the population, especially those of lower socioeconomic status and lower educational attainment, which are often the groups at highest risk.
We’re not anywhere close to a peak, as sensors continue to get better, form factors are changing, battery life is getting better. We’re really still on the flat part of the curve of what’s possible through digital health.
Sensors are a big part of it, but the other two, equally important pillars of digital health sensors – taking advantage of machine learning and 24/7 connectivity – are going to be equally critically important to the successful expansion of digital health solutions. The latter, which health care has been slowest to address, is the communication aspect, the 24/7 connectivity, returning results to participants, etc. Part of precision medicine also has to be precision communication.