Mike Snyder at Stanford symposium on the potential of smartphones to better understand diseases, inc ME/CFS (Simon M blog)

Mike Snyder is a pioneer and I thought he gave the standout talk at Stanford this year.

Stanford symposium: the potential of smartphones to better understand diseases, including ME/CFS

“We think smartphones will be the most important health tool in future”, said Dr Michael Snyder at the recent Stanford ME/CFS symposium. Snyder heads up the Stanford Centre for Genomics and Precision Medicine, as well as the Genetics Department at Stanford University.

Dr Michael Snyder. Check out all those wearables on his wrists.​

His presentation highlighted the potential of technologies to monitor health and better understand diseases – including ME/CFS.

Snyder’s work on personal, data-driven medicine kicked off with a landmark paper in 2012 (which wasn’t covered in his talk). This revealed insights from deep molecular profiling over 14 months of a single person – Snyder himself.

The study tracked over 40,000 different biological molecules and integrated the findings with data from clinical tests and other sources. This is in addition to analysing Snyder’s full genome sequence.

From data to diagnosis
The study team’s analysis uncovered extensive changes in biological pathways over time, both when Snyder was healthy and when he was sick. But the standout finding was spotting diabetes as it developed in Snyder.

Snyder had no family history of diabetes, nor any lifestyle factor such as carrying too much weight. Even so, the genome analysis indicated he had a genetic risk of type II diabetes. And after one viral infection, his blood glucose levels ramped up to much higher levels than before the infection: it looked like he might have developed by diabetes. Snyder had no obvious signs of the illness, but a visit to the doctor confirmed that he did indeed now have type II diabetes.

Prompt detection of the disease the moment it developed gave Snyder the chance to act fast. And he has been able to manage his illness through dietary and lifestyle changes, and without needing any medication. LINK

A second study showed this early detection was no one-off. Taking a similar approach, Snyder and his team made over 50 discoveries of significant health problems, including heart problems, a blood cancer and two precancerous conditions, among the 109 people studied. Most cases involved early detection of disease before symptoms had even started.

Taking it to the masses: the shift to wearable devices
The results were impressive but depended on a large budget and access to a state-of-the-art lab. Snyder wondered if a simpler approach would work.

The expanded study had included data from wearables, such as Fitbits, tracking a range of factors including physical activity, heart rate and skin temperature. Could wearables alone provide useful information about health?

So, Snyder began tracking himself with wearables. At one point in his talk, he held up his arms to show he was manacled with wearables devices. He said that even his ring was actually a biosensor. To top it all, he produced a radiation monitor from his pocket.

Read the full blog

 

Wearable devices certainly can play an interesting role in management, I'm not so sure about "early detection" though, I'm reminded of overdiagnosis and false positives due to cancer screening for example.

 

I'm very interested in this, and am trying to process my downloaded FitBit data to look for signs of this current relapse. Struggling a bit with processing the JSON datafile format that they use and getting it into Python though. I've not done any programming for years, and my experience was all with Fortran. Times have moved on! I'm finding the learning curve a bit too steep.
The fact that he's working with Ron Davis makes me happy.

 

Yeah, the more I thought about @Simon M 's post, the more it made me think of all the weird data my body gives off daily. Neither me nor my doctor would know what to do with that information. Indeed that has been the conclusion of I've come to with regard to the little bit of data I have collected over the years -- what the hell do I do with all of this? Without a computational biologist handy, there's been no discernible pattern to emerge, aside from the stuff that suggests I should be dead already.

 

The overdiagnosis problem is because they look for indications not the actual disease. These sort of systems may actually prevent some of these problems.

Of course the "downside" of breast cancer screening is often given as "women get themselves upset by being recalled" so the "woman as neurotic" is made much of when simple education would work just as well.

Ringding I use this program to access my fitbit heart rate data. http://gondwanasoftware.net.au/exerbyte/

It downloads as a .csv file and I convert it to .ods. Your data is graphed on site so you can have a quick check. It gives the data without the 5 minute average on the fitbit site. I have used it for almost 2 years now.

 

Thanks for all the comments

On the blog and the facebook page about it, there are quite a lot of people saying that they already use resting heart rate to help avoid PEM eg

Important points.

It's worth noting that all Snyder's health "discoveries" were checked out with clinicians (avoiding the false positives issue) - the idea is that their smart programming homes in on the relevant data and indicates what needs checking out. If that's cancer or diabetes that's good - early intervention can save/improve lives. Certainly in the UK, the lack of early diagnosis is the biggest single reason that survival rates here are relatively poor. Too many die here because of late diagnosis. (Unnecessary treatment is an important but separate issue.)

Snyder stressed that the aim is a tool that shows patients and doctors what's relevant.

 

@Mithriel thanks, I wasn't aware of that one. I've used something similar on github for intraday HR. What I'm trying to do at the moment is look at daily measurements, RHR, time in different HR zones, steps etc. as I have data from early on in my recovery from when I was first ill, to being 80% of normal, to this current relapse. I'm interested in whether there's anyway of slicing the data that could have given me a warning, as this came out of nowhere. It might be a hiding to nothing but at least it's giving me a way of exercising the little grey cells. It's three years worth data so it could be interesting.

 

@Simon M using RHR is exactly what I had been doing before this relapse. Not as the only guide but as one of the main ones. For some reason I crashed badly so am interested in if (with the limited other data I have) I can find any other signs. I hope to recover enough to get back to work part time and realise I'll have to be very careful if I do so. Any tools that could help me avoid a repeat would be great.

 

@Simon M thanks for highlighting you've a Facebook page. Now following.