WEBVTT - Engineering Your Own Pancreas (Rebroadcast) 0:00:00.240 --> 0:00:04.720 High Prognosis listeners, We're looking back at our favorite episodes 0:00:04.760 --> 0:00:08.280 from previous seasons over the next four weeks. We start 0:00:08.320 --> 0:00:11.640 with our very first episode ever, where a community came 0:00:11.640 --> 0:00:15.640 together to develop a biohack that revolutionized life for diabetics. 0:00:16.360 --> 0:00:24.400 Thanks for listening and enjoy the show. Hm hm hm. 0:00:28.480 --> 0:00:30.360 What do you do when a piece of your body's 0:00:30.400 --> 0:00:34.760 hardware goes haywire? If you have Type one diabetes, that 0:00:34.920 --> 0:00:38.440 damaged hardware is your pancreas. There's no way to fix 0:00:38.440 --> 0:00:41.280 a broken pancreas, but you can build a new one. 0:00:43.120 --> 0:00:51.320 Welcome to Prognosis. I'm your host, Michelle fake Cortez. For 0:00:51.360 --> 0:00:54.600 the next eight weeks, we'll be exploring science and technology 0:00:54.680 --> 0:00:57.800 and medical research. We'll look at how the treatments of 0:00:57.840 --> 0:01:00.680 tomorrow get into the hands of patients who need them 0:01:00.720 --> 0:01:03.720 the most. Along the way, we'll meet people living on 0:01:03.760 --> 0:01:09.399 the edge of innovation. There's the scientist genetically engineering frogs 0:01:09.400 --> 0:01:13.480 in his garage, drug company researchers using newly unlocked secrets 0:01:13.480 --> 0:01:16.840 of the body to cure deadly diseases. And we'll hear 0:01:16.840 --> 0:01:19.000 about one man's quest in the desert to turn a 0:01:19.040 --> 0:01:23.480 psychedelic drug into a treatment for depression. We begin our 0:01:23.520 --> 0:01:25.640 journey with the do it yourself device that replaces one 0:01:25.680 --> 0:01:39.039 of the body's most critical organs. One point three million 0:01:39.080 --> 0:01:42.720 Americans have Type one diabetes. That's the most severe kind. 0:01:43.280 --> 0:01:45.720 It develops one cells in the pancreas that make insulin 0:01:45.959 --> 0:01:49.920 stop working. Why it happens is still a mystery. Insulin 0:01:50.000 --> 0:01:53.440 is the hormone that converts food into energy. A healthy 0:01:53.440 --> 0:01:56.960 pancreas is a marvel pumping out precise amounts of insulin 0:01:57.080 --> 0:02:00.720 at just the right time. They're dozens of things that 0:02:00.760 --> 0:02:05.040 affect how much is made, eating, drinking, exercising, sleeping, you 0:02:05.160 --> 0:02:07.760 name it. It's a problem. If you don't make any 0:02:07.880 --> 0:02:09.960 sugar builds up in the blood. It comes up your 0:02:10.040 --> 0:02:12.519 organs like the kidneys in the heart and can cause 0:02:12.560 --> 0:02:16.320 an early death. Diabetics can inject replacement insulin, but they 0:02:16.360 --> 0:02:18.639 have to figure out how much to give and when. 0:02:19.400 --> 0:02:22.800 Answering those two seemingly simple questions is what makes life 0:02:22.800 --> 0:02:26.040 with diabetes miserable. You have to take on the duties 0:02:26.040 --> 0:02:30.080 of a healthy pancreas all day, all night, three and 0:02:30.080 --> 0:02:32.959 sixty five days of the year. There are tools to 0:02:33.040 --> 0:02:36.680 help glucose monitors track blood sugar levels, so people can 0:02:36.720 --> 0:02:39.720 calculate how much insulin they need, and there are pumps 0:02:39.720 --> 0:02:42.280 to deliver the insulin, but you have to tell it 0:02:42.320 --> 0:02:45.600 what to do. Americans spend more than eight billion dollars 0:02:45.639 --> 0:02:48.200 a year on these devices, but no one's been able 0:02:48.240 --> 0:02:50.680 to get them connected and working together to deliver the 0:02:50.720 --> 0:02:53.239 right amount of insulin the way a real pancreas does 0:02:55.200 --> 0:02:59.000 until now. This is a story about strangers coming together 0:02:59.080 --> 0:03:02.240 from all over the world, driven by a desperate need. 0:03:02.760 --> 0:03:05.639 They banded together on social media to take matters into 0:03:05.639 --> 0:03:09.160 their own hands. Now, just to be clear, these folks 0:03:09.200 --> 0:03:12.359 didn't build a mechanical version of the pancreas. What they're 0:03:12.400 --> 0:03:15.840 actually doing is linking their insulin pumps to their glucost 0:03:15.840 --> 0:03:19.960 monitors and getting them to work together without any human involvement. 0:03:20.600 --> 0:03:24.320 None of this is FDA approved. Still, thousands of people 0:03:24.320 --> 0:03:27.519 have tried it out on themselves and on their diabetic kids, 0:03:28.160 --> 0:03:31.440 and they love it. More are signing up every day. 0:03:31.840 --> 0:03:43.880 Here's Bloomberg's Naomi Kresky with the story. Building your own 0:03:43.920 --> 0:03:48.680 pancreas can be pretty intimidating. There's software code to install, 0:03:49.000 --> 0:03:52.680 and bluetooth and radio signals to sink. Among the four 0:03:52.840 --> 0:03:56.680 electronic devices that you'll need to link together, but there 0:03:56.720 --> 0:03:59.840 will be lots of people to help. Someone you're likely 0:03:59.840 --> 0:04:03.120 to come across early on is a mom from outside 0:04:03.120 --> 0:04:06.839 Toronto named Kate Farnsworth. Some people call her Saying Kate 0:04:07.040 --> 0:04:09.600 because she spent so much time helping people build their 0:04:09.600 --> 0:04:15.120 own diabetes technology. Kate lives in a quiet suburb about 0:04:15.120 --> 0:04:19.000 an hour northwest of Toronto. She realized that something was 0:04:19.240 --> 0:04:21.960 up with her daughter, Sydney more than six years ago. 0:04:22.640 --> 0:04:25.880 Sydney had started drinking a lot of water and eating 0:04:25.960 --> 0:04:30.599 a lot more than usual. I used to say she 0:04:30.720 --> 0:04:32.880 was eating like a teenage boy, but she was an 0:04:32.880 --> 0:04:35.800 eight year old girl. We used to have to pack 0:04:35.839 --> 0:04:40.479 an extra snack after school and leaders of extra water, 0:04:40.560 --> 0:04:43.599 and she'd cry when she ran out of water. But 0:04:43.760 --> 0:04:46.920 even with all that extra eating, Sydney was losing weight. 0:04:47.560 --> 0:04:52.200 Kate googled her symptoms and diabetes popped up. So I 0:04:52.279 --> 0:04:56.080 took her to the family doctor, and the family doctor said, well, 0:04:56.080 --> 0:04:58.560 that's an easy test. We can just poke her finger 0:04:58.680 --> 0:05:01.600 and see what's going on. Don't worry about it. And 0:05:01.640 --> 0:05:04.720 the nurse poked her finger and she turned to the 0:05:04.720 --> 0:05:08.080 other nurse and she said it's not working. It's not 0:05:08.240 --> 0:05:11.440 It doesn't say a number, it just says high. And 0:05:11.520 --> 0:05:14.520 the nurse said that's because her blood sugar is too 0:05:14.600 --> 0:05:17.240 high for the meter to read. And that was our 0:05:17.800 --> 0:05:23.080 introduction to diabetes. Sydney is now fifteen years old. She 0:05:23.200 --> 0:05:25.800 still remembers how tough things were when she first got 0:05:25.839 --> 0:05:32.200 the diagnosis. I was terrified because that's like I was eight. 0:05:32.279 --> 0:05:35.800 I didn't want anything like this happening, and it was 0:05:35.880 --> 0:05:41.120 just there was life changing. Yeah, people with type one 0:05:41.200 --> 0:05:45.440 diabetes need to take replacement insulin, but messing at the 0:05:45.480 --> 0:05:49.760 dosage can have catastrophic results, including going into a coma 0:05:49.920 --> 0:05:53.120 or even dying, and figuring out how much to take 0:05:53.279 --> 0:05:56.919 is really complicated. There are a ton of things that 0:05:57.000 --> 0:06:02.920 affect blood sugar. So sleep can affect blood sugar, anxiety, adrenaline, 0:06:03.400 --> 0:06:11.080 food stress, um, exercise, but just basically everything affects blood sugar. 0:06:11.279 --> 0:06:15.400 Roller coasters make her blood sugar drop, walking to school 0:06:15.480 --> 0:06:19.000 makes her blood sugar drop. I've heard people say that 0:06:20.040 --> 0:06:23.040 boys walking by make their daughter's blood sugar go up 0:06:23.040 --> 0:06:26.720 as well. I have not noticed that myself, but I'm 0:06:26.760 --> 0:06:30.200 not paying that close attention. So Kate and her husband, Dave. 0:06:30.360 --> 0:06:33.159 We're getting up several times a night to check Sydney's 0:06:33.160 --> 0:06:36.440 blood sugar. This is something that I heard a lot 0:06:36.560 --> 0:06:40.280 from parents of kids with diabetes. When your kids are asleep, 0:06:40.400 --> 0:06:43.400 it's hard for them to know when blood sugar levels 0:06:43.440 --> 0:06:46.679 are dropping too low. There are these little devices called 0:06:46.680 --> 0:06:51.080 continuous glucose monitors or cgms for short, that sound and 0:06:51.160 --> 0:06:53.720 alarm when their readings are dipping too low. But they 0:06:53.760 --> 0:06:56.520 can be really hard to hear, and until recently, you 0:06:56.560 --> 0:06:59.200 couldn't tune into them from even a few rooms away. 0:06:59.680 --> 0:07:05.840 They're also expensive. At first, Sydney didn't even have one, 0:07:07.040 --> 0:07:10.520 so Davi would check Sydney's blood sugar at midnight before 0:07:11.200 --> 0:07:14.520 he went to bed. Kate would set her alarm for 0:07:14.600 --> 0:07:18.160 three am, when she set another alarm for six am. 0:07:18.240 --> 0:07:21.600 They were so tired all the time. Kate thought there 0:07:21.640 --> 0:07:28.280 had to be a better way. One day, I was googling, um, 0:07:28.320 --> 0:07:32.760 researching whatever you want to call it, all things diabetes, 0:07:32.800 --> 0:07:35.480 and I came across a blog and it had a 0:07:35.520 --> 0:07:39.080 little girl who had a backpack with a cell phone 0:07:39.280 --> 0:07:43.320 connected to this receiver Sydney wears, and it was broadcasting 0:07:43.360 --> 0:07:46.960 their blood sugars to their parents, and I thought, that's 0:07:47.000 --> 0:07:48.960 that's what I need. I need to be able to 0:07:48.960 --> 0:07:51.840 see her blood sugar when she's not with me. Kate 0:07:51.960 --> 0:07:54.600 found a group of parents who had managed to pull 0:07:54.720 --> 0:07:59.800 data from signs onto their smartphones and watches. This group 0:07:59.880 --> 0:08:02.280 is called Knight Scout, and it came up with the 0:08:02.400 --> 0:08:06.600 first big d I Y advance. It helped parents track 0:08:07.080 --> 0:08:10.040 how their kids were doing at school, at night, or 0:08:10.080 --> 0:08:15.120 any other times when they weren't together. Up until that time, 0:08:15.680 --> 0:08:20.800 you had to be within fifteen feet of the continuous 0:08:20.800 --> 0:08:26.480 glucose monitor receiver that you received from uh dex commer Metronic. 0:08:26.960 --> 0:08:29.920 This allowed you to be anywhere in the world and 0:08:30.080 --> 0:08:34.280 still see your blood glucose values. That's Weston Nordgren, a 0:08:34.400 --> 0:08:38.400 dad from Bakersfield, California. He helped start Night Scout in 0:08:38.440 --> 0:08:43.720 a The group operates on the principle of pay it forward. 0:08:44.400 --> 0:08:46.800 Every single family that set up the d I Y 0:08:46.880 --> 0:08:50.880 system was supposed to help someone else figure it out too, 0:08:51.240 --> 0:08:54.560 So they started with five families, and what happened after 0:08:54.640 --> 0:09:01.000 that defied everyone's expectations. Within six weeks of creating that group, 0:09:01.720 --> 0:09:09.040 there were one thousand, seven hundred families. By September, there 0:09:09.040 --> 0:09:13.920 were seven thousand families. By December there were ten thousand families. 0:09:14.640 --> 0:09:19.880 So it just exploded. So now parents could watch their 0:09:19.960 --> 0:09:23.439 kids blood glucose. That was a big deal, but some 0:09:23.480 --> 0:09:26.839 people were already thinking of taking it further. The next 0:09:26.840 --> 0:09:30.479 advance would have been impossible if not for a blunder 0:09:30.480 --> 0:09:35.240 by Metronic, the world's biggest medical device maker, left a 0:09:35.280 --> 0:09:39.079 security gap and some of its insulin pumps the wireless 0:09:39.200 --> 0:09:43.160 radio frequency link was open to hackers. Ben West, a 0:09:43.200 --> 0:09:47.120 programmer and diabetes patient in California, found the open back 0:09:47.160 --> 0:09:50.440 door and walked right through. I actually don't describe it 0:09:50.480 --> 0:09:53.959 as a backdoor. Um, it's more like the front door 0:09:54.120 --> 0:09:57.600 wasn't put on, it wasn't locked, or it wasn't clearly no, 0:09:57.800 --> 0:10:00.160 Like there's just no door. There's just there's just no door. 0:10:02.000 --> 0:10:05.680 What door. That explains to me that there are two 0:10:05.720 --> 0:10:08.600 ways to protect against a hack. You can make a 0:10:08.640 --> 0:10:12.559 strong defense or you can just hide it. Try to 0:10:12.640 --> 0:10:16.120 keep the soft and vulnerable spot a secret so it's safe. 0:10:16.720 --> 0:10:19.800 And that's the choice the Metronic made when they designed 0:10:19.840 --> 0:10:23.000 this device. We're just gonna leave it so that it's 0:10:23.040 --> 0:10:25.120 easy for us to test and get out the door 0:10:25.400 --> 0:10:28.760 and ship very quickly. We'll keep how it works a secret, 0:10:28.800 --> 0:10:31.720 and so we're claiming that no one will know how 0:10:31.720 --> 0:10:33.600 it works, and since no one knows how it works, 0:10:33.600 --> 0:10:35.960 no one will ever attack it. That that's basically the 0:10:36.000 --> 0:10:39.800 tactic piptok. Now, it turns out that you know, I 0:10:39.880 --> 0:10:42.240 came along and I said, well, it's on my body, 0:10:42.800 --> 0:10:45.200 and I actually need to know how it works in 0:10:45.320 --> 0:10:47.280 order to make sure that it's not doing the wrong 0:10:47.320 --> 0:10:50.760 thing to my body. Ben started the project in winter 0:10:51.040 --> 0:10:53.880 of two thousand nine. He was already working at that 0:10:53.920 --> 0:10:56.800 point as a software engineer, and then he got a 0:10:56.880 --> 0:11:00.240 job at a cloud computing startup. He didn't have whole 0:11:00.320 --> 0:11:03.200 lot of free time, so he would take his metronic 0:11:03.280 --> 0:11:06.240 pump home with him to his family over Christmas and 0:11:06.320 --> 0:11:09.520 New Years, and we would spend the you know, almost 0:11:09.559 --> 0:11:12.480 the entire holiday just working on this thing. You know. 0:11:12.480 --> 0:11:15.280 We would take poster paper and start trying to figure 0:11:15.280 --> 0:11:18.800 out individual bits and bites and the meaning of individual codes, 0:11:19.360 --> 0:11:21.400 and we would try to dissect, you know, what are 0:11:21.400 --> 0:11:24.880 these time stamps and how are they encoding times and dates? 0:11:24.960 --> 0:11:29.280 And we would kind of um pick these different topical 0:11:29.320 --> 0:11:32.600 areas and really intensely attack it over the holiday period, 0:11:32.920 --> 0:11:36.880 and those were usually very very productive sessions. What what 0:11:36.920 --> 0:11:39.800 we would do is kind of form necklaces um out 0:11:39.800 --> 0:11:42.640 of these things, so that you know, the glucoast meter 0:11:42.800 --> 0:11:45.800 plugs into a USB cable, and the USB cable goes 0:11:45.840 --> 0:11:49.679 into a computer like a gumsticks or a Raspberry Pie, 0:11:50.040 --> 0:11:52.599 and then that goes back into another USB cable that 0:11:52.640 --> 0:11:54.920 goes into another thing. So pretty soon you've got this 0:11:55.000 --> 0:11:57.560 thing that kind of twists itself around and loops around. 0:11:57.800 --> 0:12:01.040 And we would kind of wear those around so that 0:12:01.720 --> 0:12:05.079 so that we could continue working while moving around the house, 0:12:05.160 --> 0:12:10.160 I suppose. So after about five years of work, Ben 0:12:10.200 --> 0:12:14.920 managed to reverse engineer the pumps communications code. That means 0:12:15.080 --> 0:12:17.319 he figured out how to tell the device what to do. 0:12:18.760 --> 0:12:21.800 Now that d I Wires had two pieces of the puzzle, 0:12:22.000 --> 0:12:25.240 they would need to make an artificial pancreasis. They had 0:12:25.360 --> 0:12:28.480 night Scout, which was pulling the blood glucos data off 0:12:28.480 --> 0:12:32.240 glucose monitors for Sydney and Kate. That meant either of 0:12:32.280 --> 0:12:34.880 them could watch Sydney's blood sugar from a smart watch 0:12:35.040 --> 0:12:37.559 while Sydney was in school or if she was asleep. 0:12:38.480 --> 0:12:42.200 Now Ben had hacked the pump. The final piece of 0:12:42.240 --> 0:12:46.120 the puzzle would be an algorithm. That's a script that 0:12:46.200 --> 0:12:50.440 shows a computer how to do a repetitive but complicated 0:12:50.480 --> 0:13:00.000 task like calculating insulin doses. Entered Dana Lewis. She lived 0:13:00.000 --> 0:13:03.360 since Seattle, and she has diabetes. So if I sat 0:13:03.400 --> 0:13:05.520 down to have lunch, I would be looking at a 0:13:05.559 --> 0:13:07.400 plate of food and I would have to do math 0:13:07.440 --> 0:13:09.960 to figure out how many grams of carbohydrate around the food. 0:13:10.400 --> 0:13:12.400 I'd also have to look at what my current blood 0:13:12.400 --> 0:13:14.800 glucose is, and if I had a c GM, I'd 0:13:14.840 --> 0:13:17.120 be looking at the past couple of hours trying to 0:13:17.200 --> 0:13:19.400 visually predict in my head what is the trend of 0:13:19.400 --> 0:13:21.839 my blood suger what's going to happen. I also had 0:13:21.840 --> 0:13:24.320 to look at my pump and get the information about 0:13:24.360 --> 0:13:26.840 how much insulin I already had active in my body, 0:13:27.400 --> 0:13:30.480 and so I had to combine the current glucose kind 0:13:30.480 --> 0:13:32.880 of what has happened in the past, this insulin amount, 0:13:33.160 --> 0:13:35.320 predict what I was going to happen, add in the 0:13:35.360 --> 0:13:37.439 calculation of how much food and how that was going 0:13:37.480 --> 0:13:40.320 to impact my body, and decide how much insulin do 0:13:40.360 --> 0:13:42.760 I need? And not only how much insulin do I need, 0:13:42.800 --> 0:13:45.600 but how should I time it? Dana and her partner, 0:13:45.679 --> 0:13:49.160 Scott Liebrand, decided that it would be a lot simpler 0:13:49.200 --> 0:13:52.600 to let the computer do some of that calculating, so 0:13:52.960 --> 0:13:56.559 they wrote an algorithm. At first, it would only spit 0:13:56.559 --> 0:14:00.679 out recommendations. Then they took their project to a conference 0:14:01.080 --> 0:14:05.319 and they met Ben. It was June, and we stood 0:14:05.320 --> 0:14:07.320 beside him kind of demonstrating what we've been working on, 0:14:07.360 --> 0:14:09.560 and he was demonstrating the fact that he could read 0:14:09.600 --> 0:14:12.840 and write to this metronic pump. And it actually wasn't 0:14:12.880 --> 0:14:15.880 until a few weeks or months later that another light 0:14:15.880 --> 0:14:18.160 bulb went off and we went, hey, we've got this 0:14:18.200 --> 0:14:21.640 algorithm that takes data and predicts into the future. What 0:14:21.680 --> 0:14:25.120 should happen if we combined it with his code to 0:14:25.320 --> 0:14:28.240 talk between the devices. Instead of telling the human what 0:14:28.320 --> 0:14:31.080 to do, we could actually write directly to the insulin 0:14:31.120 --> 0:14:33.760 pump and tell it to give more or less insulin. 0:14:34.320 --> 0:14:36.760 So then they had to figure out how to close 0:14:36.800 --> 0:14:40.160 the loop. They didn't start out knowing exactly what they 0:14:40.160 --> 0:14:43.840 would build, but they just started working on the problem 0:14:44.120 --> 0:14:48.120 step by step, trading the ideas back and forth. So 0:14:48.240 --> 0:14:50.480 I live in Seattle. Scott, who was then my boyfriend, 0:14:50.560 --> 0:14:53.200 lived twenty miles away in Everett, Washington, and Ben was 0:14:53.200 --> 0:14:56.280 based out of San Francisco, and so we would collaborate 0:14:56.320 --> 0:14:58.080 in a variety of ways. We would use a chat 0:14:58.120 --> 0:15:00.480 channel called getter, which is connected to at hub, where 0:15:00.480 --> 0:15:04.400 we would get on and start talking. We used Twitter, UM. 0:15:04.440 --> 0:15:06.280 We also just used phone calls where we would get 0:15:06.280 --> 0:15:08.800 on speaker phone and be programming and using the chat 0:15:08.880 --> 0:15:10.920 channel and also talking by phone. There was also times 0:15:10.920 --> 0:15:12.760 where Scott would need to go down to San Francisco 0:15:12.800 --> 0:15:14.640 for work, so he and Ben would meet up. Sometimes 0:15:14.640 --> 0:15:16.680 I would be in San Francisco. We beat up in person, 0:15:17.120 --> 0:15:19.800 you know, on nights and weekends, um, and really just 0:15:20.040 --> 0:15:22.600 kind of plucking away at Oh, we found a problem, 0:15:22.640 --> 0:15:24.640 let's resolve it. Okay, we found the next problem in 0:15:24.680 --> 0:15:27.320 the next obstacle, let's figure out how to tackle it. 0:15:27.680 --> 0:15:31.080 In December, the group finally felt like they had surmounted 0:15:31.160 --> 0:15:35.040 all of the obstacles they enough finally felt confident enough 0:15:35.120 --> 0:15:39.000 to become the first human tester for her new artificial pancreas. 0:15:39.600 --> 0:15:41.960 Originally my plan was just to test it for a 0:15:41.960 --> 0:15:43.920 couple of hours, but I was like, you know what, 0:15:44.360 --> 0:15:46.280 I really I want to go to sleep and I 0:15:46.280 --> 0:15:48.600 want to go to sleep with this thing working. And 0:15:48.680 --> 0:15:51.320 so we set up all of our backup alarms, and 0:15:51.440 --> 0:15:53.800 I went off and went to sleep with the system. 0:15:53.840 --> 0:15:55.920 And I woke up the next morning and I felt 0:15:56.280 --> 0:15:58.840 better than I ever had before, because I woke up 0:15:58.840 --> 0:16:01.480 with my blood sugar in perfect range where I wanted 0:16:01.480 --> 0:16:04.400 to be. I hadn't been low overnight, my bletcher hadn't 0:16:04.440 --> 0:16:06.560 been high overnight. I wasn't interrupted in my sleep. I 0:16:06.600 --> 0:16:08.160 didn't have to drink a juice box, I didn't have 0:16:08.160 --> 0:16:10.440 to press a button on my pump. I got ten 0:16:10.800 --> 0:16:14.720 uninterrupted hours of sleep, and it was glorious. The d 0:16:14.840 --> 0:16:18.480 i Y community was all about open source, or sharing code, 0:16:19.040 --> 0:16:21.960 so Scott, Dana, and Ben wanted to share what they 0:16:22.120 --> 0:16:24.840 learned so that others could also feel and sleep better. 0:16:25.520 --> 0:16:28.400 But it wasn't just up to them. There's also the 0:16:28.440 --> 0:16:31.360 f d A, or the Food and Drug Administration, which 0:16:31.360 --> 0:16:34.680 regulates devices in the US. They were worried that the 0:16:34.720 --> 0:16:37.400 FDA would come down on them for sharing something they'd 0:16:37.400 --> 0:16:41.040 cooked up at home. In the end, after talking to 0:16:41.040 --> 0:16:43.680 the f d A weighing the pros and cons, they 0:16:43.680 --> 0:16:47.320 decided to post their work online, with one important caveat. 0:16:47.880 --> 0:16:50.280 The people who wanted to use it would have to 0:16:50.320 --> 0:16:53.920 build it themselves first. Only if you tried it out 0:16:54.760 --> 0:16:58.440 then more and more. Kay was watching too. She was 0:16:58.480 --> 0:17:03.200 intrigued but also intimidated. It seemed very complicated to build. 0:17:03.960 --> 0:17:06.880 But then another member of the community designed a more 0:17:07.000 --> 0:17:11.480 user friendly version for the iPhone called loop In. Kate 0:17:11.560 --> 0:17:14.679 saw it in action. A friend's little boy was running 0:17:14.720 --> 0:17:19.199 around having his blood sugar automatically adjusted by the system. 0:17:19.200 --> 0:17:22.120 She started looking for the parts she needed. That same day, 0:17:22.560 --> 0:17:27.280 Medtronic had closed Ben's open door on its newer pumps, 0:17:27.800 --> 0:17:30.320 so Kate had to get a hold of an older model. 0:17:30.840 --> 0:17:34.800 Someone gave her one. Then she ordered a tiny Bluetooth 0:17:34.800 --> 0:17:38.919 equipped computer which enabled the pieces of the Loop system, 0:17:38.960 --> 0:17:42.520 so the insulin pump, the glucose monitor, the iPhone app 0:17:42.560 --> 0:17:45.720 with the algorithm to talk to each other another d 0:17:45.880 --> 0:17:49.520 I wire. A dad from Minnesota named Peach Farm had 0:17:49.560 --> 0:17:52.400 designed the device for his daughter and named it after her, 0:17:52.760 --> 0:18:09.800 calling it the Riley Link. Kate got to work. There 0:18:09.840 --> 0:18:13.280 weren't a lot of instructions to follow, but just enough. 0:18:14.080 --> 0:18:17.280 I built the app on her phone and then I 0:18:17.359 --> 0:18:21.960 set up um the spare pump and the Riley link 0:18:22.200 --> 0:18:26.280 and uh the iPhone and let it run, but not 0:18:26.359 --> 0:18:29.600 connected to Sydney. I just filled the insulin pump with 0:18:29.680 --> 0:18:33.200 water and let it drip out as if it were 0:18:33.280 --> 0:18:36.960 giving Sydney insulin. She watched on and off for hours 0:18:37.200 --> 0:18:40.200 as the pump pushed out tiny squirts of water into 0:18:40.240 --> 0:18:43.720 a napkin. Or Of course, it wasn't affecting her blood 0:18:43.720 --> 0:18:46.040 sugar at that time because it wasn't hooked up to her, 0:18:46.440 --> 0:18:49.080 but I could see the logic between behind what it 0:18:49.160 --> 0:18:52.480 was trying to do. Kate hooked Sydney up to the 0:18:52.520 --> 0:18:55.560 system on a Friday afternoon so she could keep a 0:18:55.600 --> 0:18:58.800 close eye on how it was working. It started keeping 0:18:58.840 --> 0:19:02.240 Sydney's blood sugar range without Kate having to do much 0:19:02.240 --> 0:19:05.960 of anything. All of a sudden. You know, life changed. 0:19:06.240 --> 0:19:10.080 When you're not spending all of your time worrying about diabetes, 0:19:10.600 --> 0:19:14.240 you have all this extra time in your life. Sydney 0:19:14.320 --> 0:19:17.879 stopped worrying too. I'm a teenager. I have my phone 0:19:17.880 --> 0:19:22.560 with me. All I have to do is open an app, 0:19:22.960 --> 0:19:27.000 putting my carbs, get my insulin, and keep going. It's 0:19:27.040 --> 0:19:30.320 just so much easier than having to go a whole 0:19:30.320 --> 0:19:33.439 Another level of just pulling out my pump, doing all that, 0:19:33.560 --> 0:19:36.080 clipping it back on, going back to what I was doing. 0:19:36.400 --> 0:19:39.359 Does anybody even really have to know if you're adjusting 0:19:40.119 --> 0:19:43.080 your carbs now? I mean, would anybody even realize what 0:19:43.119 --> 0:19:46.240 you're doing? No. Sometimes I'm sitting at lunch and I'm 0:19:46.280 --> 0:19:48.399 just giving insult on my phone and my friends just 0:19:48.560 --> 0:19:51.959 think I'm on my phone looking at something just like 0:19:52.119 --> 0:20:02.040 totally normal. Yeah, just to be clear, this is a 0:20:02.160 --> 0:20:05.399 d i Y hack. No pump on the commercial market 0:20:05.520 --> 0:20:09.879 lets you dose insulin from your iPhone. Yet Metronic, the 0:20:09.920 --> 0:20:13.480 world's biggest maker of insulin pumps, is watching closely. They're 0:20:13.480 --> 0:20:17.280 working on an artificial pancrease too. They have one device 0:20:17.359 --> 0:20:21.359 on the market now, but it's not as advanced. It 0:20:21.440 --> 0:20:24.800 doesn't adjust insulin for meal times like Dana System does, 0:20:25.520 --> 0:20:28.760 and it doesn't allow dosing from an iPhone like Sydney's does. 0:20:29.160 --> 0:20:32.320 What it does do is make the tiny little adjustments 0:20:32.359 --> 0:20:36.159 that happen in between meals. That's already enough to be 0:20:36.280 --> 0:20:41.040 a big advance. The device, which sells for seven thousand dollars, 0:20:41.680 --> 0:20:45.760 is going like hotcakes. So far more than one thousand 0:20:45.760 --> 0:20:49.680 people have one. But there's another big difference between d 0:20:49.720 --> 0:20:53.440 I Y systems and the equipment Metronics cells. Before its 0:20:53.440 --> 0:20:56.840 products go on the market, Metronic needs to prove to 0:20:56.880 --> 0:20:59.440 the f d A that they will be safe for 0:20:59.520 --> 0:21:03.160 every one of the thousands or sometimes millions of people 0:21:03.200 --> 0:21:05.760 who buy them. I talked to Ali d and Natty, 0:21:06.080 --> 0:21:10.000 who handles R and D in Metronics. Intensive Insulin Management 0:21:10.080 --> 0:21:14.119 Unit were regulated for a reason, to make sure that 0:21:14.160 --> 0:21:17.960 we control as many of those variables as we can 0:21:18.080 --> 0:21:21.440 and think about all the potential corner cases so that 0:21:21.480 --> 0:21:25.919 you know, any something catastrophic doesn't happen, because ultimately, these 0:21:26.359 --> 0:21:30.359 patients and families are dosing insulin, and insulin can be 0:21:30.400 --> 0:21:34.639 a very very dangerous thing. It's hard to track whether 0:21:34.720 --> 0:21:37.480 anyone was hurt while using a d I Y system. 0:21:37.880 --> 0:21:40.680 Everyone I met while reporting the story said they had 0:21:40.800 --> 0:21:45.040 never heard of anybody being seriously injured. But diabetes is 0:21:45.040 --> 0:21:49.760 a dangerous disease and the risk does exist. Some of 0:21:49.760 --> 0:21:52.480 the d I wires are working towards the same f 0:21:52.640 --> 0:21:56.320 d A seal of approval that Metronic has. Last month, 0:21:56.760 --> 0:22:00.199 an open source group called tide Pool announced that is 0:22:00.200 --> 0:22:04.000 building an FDA regulated version of Loop, which is the 0:22:04.040 --> 0:22:06.960 system that Sydney is using. The goal is to have 0:22:07.040 --> 0:22:10.680 an official product that's available on the iPhone app store 0:22:11.160 --> 0:22:14.760 and compatible with pumps that you can buy today. If 0:22:14.760 --> 0:22:18.760 they succeed, what was once a radical approach to diabetes 0:22:18.920 --> 0:22:26.000 will become perhaps mainstream. For now, though. Kate, Sydney, and 0:22:26.080 --> 0:22:28.360 the other d I wires are making do with their 0:22:28.359 --> 0:22:32.800 homemade systems, and their ranks are growing. Kate took all 0:22:32.800 --> 0:22:35.720 that time that she gained from not having to micromanage 0:22:35.720 --> 0:22:39.399 Sydney's diabetes and invested it in the d I y community. 0:22:39.880 --> 0:22:43.360 She runs a Facebook group called looped that has almost 0:22:43.400 --> 0:22:47.400 six thousand members, and she spends hours every day helping 0:22:47.440 --> 0:22:52.480 other people set up systems. Like Sydneys. Kate enjoys helping others, 0:22:53.119 --> 0:22:55.720 but she's looking forward to the day when she won't 0:22:55.760 --> 0:22:58.919 have to because an FDA approved medical device will be 0:22:58.960 --> 0:23:04.160 on the market. I would love to be put out 0:23:04.160 --> 0:23:08.080 of a job one day by a company I I 0:23:08.119 --> 0:23:10.600 don't want to do this long term. I would love 0:23:10.680 --> 0:23:13.960 everyone to get a system that works for them and 0:23:14.640 --> 0:23:17.320 right out of the box, so that we don't have 0:23:17.440 --> 0:23:21.440 to keep doing this. Kate took a big risk when 0:23:21.480 --> 0:23:24.520 she plunged into building an artificial pancreas for her daughter. 0:23:25.200 --> 0:23:28.919 She didn't have medical or high tech experience. There was 0:23:28.960 --> 0:23:33.520 always the possibility that something could go terribly wrong, but 0:23:33.600 --> 0:23:37.000 the results are really hard to argue with. Sydney is 0:23:37.040 --> 0:23:39.760 in good health and she's living her life more or 0:23:39.840 --> 0:23:43.800 less like any other well adjusted teenager, and that's been 0:23:43.880 --> 0:23:48.640 worth everything. I mean, just the fact that I can 0:23:48.960 --> 0:23:52.480 live my life normally instead of having a hundred different 0:23:52.520 --> 0:23:55.800 things to focus on, and it lowers the stress a lot. 0:23:56.320 --> 0:24:01.040 Would you ever go back to the way it was before? Oh? 0:24:01.119 --> 0:24:19.760 No way, never. And that's it for this week's prognosis. 0:24:20.040 --> 0:24:23.320 Thanks for listening. Do you have a story about healthcare 0:24:23.320 --> 0:24:25.919 in the US or around the world. We want to 0:24:25.920 --> 0:24:29.120 hear from you. You can email me at m Cortes 0:24:29.200 --> 0:24:32.200 at Bloomberg dot net or find me on Twitter at 0:24:32.200 --> 0:24:35.560 fake Cortes. If you were a fan of this episode, 0:24:35.760 --> 0:24:38.159 please take a moment to rate and review us. It 0:24:38.240 --> 0:24:43.000 helps new listeners find the show. This episode was produced 0:24:43.000 --> 0:24:46.920 by Liz Smith. Our story editor was Rick Shine. Thanks 0:24:46.960 --> 0:24:51.159 to Drew Armstrong. Francesco Levie is head of Bloomberg Podcasts. 0:24:51.600 --> 0:24:52.560 We'll see you next week.