We Are Not Waiting
Dana Lewis became a reluctant self-tracker at the age of 14 when she was diagnosed with type 1 diabetes. Dana and her partner Scott Leibrand have been developing a DIY artificial pancreas that is built on top of the data flows from Dana’s continuous glucose monitor. In this talk, she describes the role that access to data plays in their DIY pancreas, with immediate and profoundly positive effects on her life. She also explains why she's constantly pushing for patients to stop waiting for solutions and go the DIY route.
Blood glucose monitor
Hello everybody and I guess that’s the problem with going on first because everybody’s going to scramble on in and they’re going to miss all the secretes I’m about to tell you, which is the fact despite my joy of being here and having found my tribe, I think somebody mentioned earlier it’s great to be here for a number of reasons.
But one of them is the fact that I joined Quantified Self tribe almost 13 years ago. I was reluctant. I was dragged kicking and screaming and I hated every moment of it, because my journey as someone becoming identified with the Quantified Self movement began with my diagnosis of type I diabetes at the age of 14.
Like a lot of people you may have heard this, but it’s really hard to empathize with until you’ve gone through your own health experience. But a data point to a person with diabetes can be the difference of life and death. And that’s what I found out.
I was testing my blood glucose 12 to 15 times per day. I was given shots, I moved to an insulin pump, and I was lucky enough to get to a continuous glucose monitor, which is a device that gives you a blood sugar data point every five minutes which is great if you happen to pick it up, press the button, look down, see the number, and you can decide whether you want to do something with it or not. It’s pretty great. And you’ve got these lines that’s saying you know you can take your high thresholds and your low thresholds, which are good. And if you cross them they’ll give you a little beep.
However, imagine you’re asleep, who here likes to sleep? I happen to love my sleep so much that my body ignores all beeps, all sounds, everything. So this thing can be going off and telling me that I am dangerously low and I will possibly have something very bad happen if I don’t pay attention to this number and I will sleep right through it. And this is a problem I knew I had for many years.
As I was a teenager in high school and living with my family, I went off to college. I moved from Alabama to Seattle without my family and lived by myself, and my biggest fear was that I would not wake up in the morning when I went to sleep, and I can’t tell you what that feeling – it’s indescribable.
If you go to sleep at night, you might be worrying about what’s on your schedule, what you’re going to wear, you know, who you’re going to see. And one of the things I always thought about was I need to text my mom and tell her I love her just in case I don’t wake up in the morning.
So I knew I had a problem. I knew it wasn’t just diabetes. I took really good care of myself, but it was the fact that this data is locked on this device, and I’ve been begging the manufacturers and saying , look, this is a really big problem. People with diabetes die in their sleep on a regular basis because they go low overnight, they don’t wake up, and they can’t do something about it whether or not somebody is in the house with them. So why don’t we make these alarms louder like you know the earthquake alarm style loud.
And of course, like wearable and all everything else, it’s a tradeoff between size, and battery, and they’ve got a bunch of other things to do, and the FDA etc. so I though you know, if I just got my data off the device I can make the alarm louder myself. I could put it on my phone, my computer, with my Q-lights, whatever I wanted to do to make it louder.
So I had this thought for years and years and years, and I finally ran into somebody on Twitter who figured out how to get the data off his sons CGM, and I thought this is great. You need to contact him. So my now fiancé and I tweeted him and said will you share your code, and thankfully he did.
So I was able to get the data off the device and make louder alarms by pushing it to my phone and making really loud sounds. It’s great. It woke me up at night. But I realized that I could do more with it now that I had freed the data from the device, and this data just didn’t belong to whoever looked at the screen now. I could give it to my family in Alabama, somebody who lived 20 miles away from me.
However, just because I was one point above the threshold or below the threshold, they don’t need to get an alarm for that, that’s okay. But if I’m 50 points below my goal threshold for 30 minutes overnight and I’m not responding, then maybe I need to know about it.
So we started by building an interface, which would alert me, but not bug the heck out of everybody else. So I had to build snooze buttons that would say, I’m awake, I’m alive, I’m doing something don’t worry. So if you do wake up and see it you’ll see I’m okay and I’m taking action.
But if I’m going to push a button, great guinea pig that I am, I might as well tell it what I’m doing and be very precise and say, I am eating this many carbohydrates, I’m taking this much insulin. So if you do an ounce of cookie, you push a button, you get this data into the system you actually get really good recommendations out of it, “This is what your blood sugar is going to be, 15 minutes, 30 minutes, hour etc.”
So you know what to do right at that moment because you have all this great data and recommendation. But you also have these future predictive real-time alerts. So that’s the system we built. It’s called DIYPS, the Do It Yourself Pancreas System. All because I wanted louder alarms, I end up getting the system and that would tell me in real time you need to pay attention if you’re not looking at the screen and pushing the button, you’re dropping drastically.
In fact, when I was walking back from the break I hadn’t looked at my CGM in the last 15 minutes, because I don’t want to look at it every 15 minutes. But I got a push notification that says, hey, you’re dropping. You need to eat something before you go on stage.
Not in quite as many words, that’s my long form interpretation, but I got a buzz that said, pay attention, take action. And I was able to take action and time for the glucose to hopefully kick in before I got up on stage that would enable me to speak a little bit slower for you guys because I know I speak fast and I only have seven minutes to talk to you guys.
So I built the system called DIYPS, that gave you real-time recommendations, but I still had to push the buttons. So I would have to wake up, look at the system, decide what to do. But we found somebody else, who built another component which enabled me to push commands to my pump. So I’m now able to use a Raspberry Pi that I bought on Amazon, this kerning stick that talks to my pump right here.
So this is a closed loop artificial pancreas, and if I asked you guys to envision what an artificial pancreas looks like, you probably envision something a lot larger and clunkier. And yes this still is a little bit too much to carry around in my small [pocket but it’s better than nothing.
And so we’re not facing the challenge of I have this amazing tool and it’s an N of one and we’ve been talking all day about how great N of one’s are. But I’m not satisfied of stopping at N of one, so we’ve created a movement called Open APS, which is an open and transparent movement to get this technology out to more people sooner.
And I think we’ll talk about this during the discussion period and you guys probably have some questions about what does FDA think. I have answers for you about what they think and what they tell me.
But our goal is to get as many people as possible to do a DIY implementation of this. I don’t want to give you this code. I don’t think that’s the safest thing, but if you have type I diabetes and you’re interested I can coach you, give you documents. It’s all posted. Go to DIYPS.org, openaps.org, read our reference design. Understand how this works.
We were talking about not left box technology, guess what? this is transparent and so are the algorithms that run on it and it’s something that anybody can understand and implement.
So one of my takeaway messages to you is the fact that don’t look at me as a patient and think not a company, we have nothing to learn. I do believe the n of one’s whether it’s somebody in academics, whether it’s a patient who’s come up with this own innovation. Whether it’s a style of treatment, whether it’s a medical device or something else, I think we have a lot to learn from them, so I love the discussion about n of one and I would like to continue that.
But my final thing to all you guts that there is this concept of this we are not waiting movement. this all started because I didn’t want to wait until three more years when the first artificial pancreas might, maybe come to market in the United States. I don’t want to die in my sleep between now and then. I don’t want my friends, loved ones, the people in the community that I come to know and love to have that same fear of going to sleep at night. And so what I’m telling you is that we as patients, we are not waiting, you’re not waiting, and I would ask you guys not to wait either and it may not be in diabetes. But diabetes was this, this is a data problem. How can we get the data off the device, what can we do with it with asthma, cancer, with heart disease, all these other things, if we just free the data and talk to patients with the ‘why’, just like what I was talking about. The ‘why’ of this is so we can live our lives. That’s what all of this is about and that’s why we’re not waiting.