The following is an interview I recently conducted for the PMWC Intl. team with Dr. Mike Snyder, Professor of Genetics at Stanford University, and in essence is a repost of the original interview that can be found on the PMWC website. The discussion with Dr. Snyder focused on why value-based healthcare is employing the power of understanding what is going on within an individual, why we need to take advantage of longitudinal data to understand a personal disease state profile, and why we need to educate physicians and insurance companies about the value and impact of various measurements that highlight deviations from a personal baseline.
Dr. Snyder is the Track 5 chair – the Health Data, Microbiome, and Patient Education track – this in addition to chairing the Health Data – Making PM Mainstream session at the upcoming PMWC 2020 Silicon Valley conference. The PMWC conference provides an unparalleled array of talks that traverse the many elements of precision medicine – new technologies, the status quo of immunotherapy, advancements in AI and machine learning, emerging therapeutics and diagnostics, the reimbursement challenges, and trends in microbiome research and applications, just to name a few.
Following is the interview with Dr. Mike Snyder.
Genomics, digital health, big data, and artificial intelligence are some of the newest technologies/fields that are reshaping medicine and healthcare.
- How will these technologies impact healthcare, and in particular individualized medicine?
- What can we realistically expect from these technologies to improve healthcare and in what timeframe?
- What is still required in order for us to take advantage of the full potential of these technologies?
MS: I believe these technologies will totally transform health care and in fact make individualized medicine a reality. And I hope it is going to be more individualized health rather than individualized medicine. Currently, we are measuring a couple of critical values in the doctor’s office and make some guesses about what the best treatment should be based on the findings gathered from these measurements. I believe in the future, we are going to be in a world where you can make tons of measurements on individuals and get lots of data about them – even when they are healthy and especially when they’re healthy. The wearables in particular are going to be powerful tools for measuring people’s parameters continuously. And so we will be monitoring people’s health both at a frequency and a data level that has never been seen before. As a consequence, we will be able to better predict disease risk, catch diseases early, and manage diseases long before they are systematic.
So yes, I believe we will be observing a transformation of healthcare and for this purpose we will need AI to enable everything. Let me share an example with you: we are in the process of a building a personalized health dashboard. We are collecting data from wearable things (which allow us to basically follow people’s trajectory), and instead of just guessing we will be able to catch something that may be wrong much earlier. By following people’s detailed trajectories and tracking deviations from their healthy baseline state, healthcare can be totally transformed. As a result, we will observe a shift in healthcare–keeping people healthy and proactive rather than reactive. We will be much more personalized and for the first time we will really take advantage of longitudinal data, which has not been really done in health care to this day.
“…the entire health care system seems totally broken. Nobody pays to keep people healthy; rather the current philosophy is to only pay to fix them and we have to change that.”
Regarding a timeframe–that is always a hard one to predict. That is to say, the wearable industry is just starting to find its way into the health market, with Apple in particular getting a lot of attention. And I assume we will see a rollout of some sort pretty quickly. Aggregating people’s data is also just starting to happen now, but to see it realized in its fullest potential will take most probably five to ten years.
We will have to educate physicians because they are more reactive and slower on the uptake. They want to first see that these new approaches truly have value. The same is true for insurance companies. And that, in fact, is the biggest challenge, in addition to understanding who is paying. This is more important right now, as the entire health care system seems totally broken. Nobody pays to keep people healthy. Rather the current philosophy is to only pay to fix them and we have to change that.
Data science is a huge driver of healthcare.
- Where do you see the biggest impact of health data in precision medicine?
- How will it help with cost reduction?
MS: On the treatment side, this is pretty much a true statement. If you get more precise treatments, for the right people, with the right drug, at the right time, then this should clearly result in cost reduction. But that is not the biggest cost savings. The biggest cost saving is training people to help individuals stay healthy. Realistically, as a general rule – if you look at the economics of these things – usually when you start adding in data to better manage people’s health, data gathering mechanisms like MRI and things, they don’t really reduce costs in the long run. But what they do is implement much, much better health. When people do get ill, that is when we see most of the cost. And most of our health care costs are from chronic diseases, especially in the last years of life.
I strongly believe that if one can avoid a chronic disease or manage it in a more precise way, it should lead to some final reduction in health costs. More than that, it is really going to just manage people’s health better. As far as who pays, that is a tricky situation. One could argue that some minimum level of care should be implemented broadly across the entire nation. Everyone should have access to some minimum level of health care. In my eyes, smart watches will get very affordable/cheap and it will be a very powerful tool for continuously measuring people’s health. This is pretty straightforward.
We may also see more innovative approaches with employers to pay for people’s health plans if they are on proactive health plans. These are the most effective health plans, which are most beneficial and will keep people healthy. Keeping their employees in a good, healthy and productive state leads to better overall company productivity. To me that is an interesting incentive model and some of these healthcare plans are innovative and can work on keeping people healthy by implementing such an incentive model.
How will data science contribute to a value-based care system?
MS: That is the number one problem. Right now we treat people based on population-based measurements. This is especially true for chronic diseases, which are usually not just one single disease but rather a combination of dozens of different diseases. Diabetes is a good example. Diabetes is a disease that in reality has 50 or more underlying diseases all lumped together by high glucose. As a result, we don’t treat diabetes patients properly at all. Yes, many do respond to Fortamet (generic: metformin) which is good. But I’m a classic example, and I am a type-2 diabetic. My blood sugar level kept going up to 7.5 mmol/L while on Fortamet. It turns out that I am actually insulin sensitive and I can produce insulin just fine. My problem is that I don’t release insulin, and so the right drug for me is Prandin (generic: repaglinide) and not Fortamet. Prandin works like a charm for me. My glucose/blood sugar level has gone down nicely over the last two months. This is a classic case of understanding the details of the underlying cause of a disease. And taking the right drug to address the underlying problem, at least in part. Value-based healthcare is employing the power of really understanding what is going on, which then results in putting the right drugs in the right people where they can be most effective. To me this is a classic case of what I would call precision diabetes. Treating the diabetes subtype properly so you can have effective results.
“Value-based healthcare is employing the power of really understanding what is going on, which then results in putting the right drugs in the right people where they can be most effective.”
As you pointed out, longitudinal profiles are very valuable for the understanding of personal disease states. The challenge with this approach is the amount of health data created, and the physician/healthcare provider not yet being able to extract insights from the data in an efficient way. What cultural shift do we need to achieve so we can share personal health data with our physicians (if they are willing to accept the data) for it to be integrated into our medical care?
MS: There are two components to this question:
- Physicians, even though they have longitudinal data, rarely look at it because it doesn’t pull up easily in their chart. It really comes down to providing the right visualization tools for data that they already have. This is a combination of technology and bringing in wearable data along with other data. We need to have the tools to collect the data, display it, and make it simple for physicians to interpret right away. It should not be that hard to have the right visualization tools to address this demand. As I mentioned earlier, we are currently in the process of building our own personal health dashboard at Q Bio.
- Physicians need to realize that these data are actually useful which is not generally the case. Not too long ago we still saw push back and heard comments such as “wearables are not very accurate.” In reality though, wearables are more accurate compared to measurements taken in a physician’s office where only one data point is measured. Getting physicians to realize that just because these measurements are different does not mean they are worse, but they are in fact better. Unfortunately, getting them to accept new technologies is a big deal.
“Individualized medicine is about longitudinal profiles with various measurements that highlight deviations from a personal baseline.”
Individualized medicine is about longitudinal profiles with various measurements that highlight deviations from a personal baseline. This can only be done if one understands what a longitudinal, healthy baseline looks like. With the right tools, we will be able to extract those key findings. We are now in the process of building prevalent smart watch tools that let individuals know when one is getting sick because there is an observation that your heart rate goes up. This simple flag can alert a person that maybe something is wrong with his / her health, especially if the individual sees no reason for an increased heart rate because they are not doing anything that should affect the heart rate.
What do we need to do to integrate and utilize longitudinal profile data?
MS: Clearly, both physicians and consumers need to be educated and to be honest, smart employers will probably want to be educated too, to keep their employees upbeat. The practical side is just building the right visualization tools for tracking and visualizing longitudinal data. It is not that hard; it just has not been done yet.
You mentioned in one of your talks that you can see “your smartphone as command center” for your own health. There are concerns about data privacy when using various healthcare apps via your smart phone.
- How can we overcome these concerns?
- What does the industry need to do to address this issue?
MS: First of all, smartphones can be just as secure as any other computer that is out there. Mine has facial recognition as an app, and therefore it does not open unless it sees me. So, I think your smart phone can be very, very secure. On the issue about data privacy in general, yes, that is certainly a big deal but it is also not a bigger deal than having your financial information kept private. People do not realize that your credit card company has a huge amount of information about everyone and as such they know where you are at all times. Mine does! If I give a talk in Sweden, they know that I am there because I paid for the plane ticket, I checked into the hotel with my credit card, and they know where and how I spent my evenings. So, your whole financial structure, which is pretty darn personal, is out there. And I would probably worry more about my financial privacy than I would about my health care privacy. Of course, I don’t mean that one should not worry about healthcare privacy. You should, especially if you care.
I personally learned a lot because of others (e.g. physicians and researchers) looking at my health information and as a result, provide good advice. It is helping me. At the broader level, we do not want to be discriminated based on healthcare information, which is why we have laws to avoid discrimination, such as based on preexisting conditions. And yes, I am a firm believer that they should really be covered. Any wealthy society should be doing that.
