There is an excellent article in the latest issue of Time magazine (Feb 9, 2015 issue) about the artificial pancreas being developed by Ed Damiano, whose son was diagnosed with Type 1 diabetes at age 11 months. My favorite part of the article was the part about the 11 year old boy who was part of the clinical trial. The article says that the boy liked the bionic pancreas so much that he ran away from the investigators conducting the test, and it took them over an hour to get the device back. A 35 year old man who tried the bionic pancreas for 22 days termed it "bliss."
I assume it's not available to non-subscribers?
I have a hard time imagining the current APs could provide more than fairly crude control with a buffer of higher-than-optimal average BGs, given the inaccuracy of current CGM technology and the lag times associated with both CGMs and current rapid-acting insulins. Fancy processing can overcome some of those shortcomings, but only to a point.
Or maybe I'm not giving the technology enough credit?
I look forward to reading about the artificial pancreas. I feel impatient waiting for it!
If it is the article "The Next Best Thing to a Cure for Diabetes" by Alexandra Sifferlin, then, yes, it appears to only be available to subscribers.
All I get is a picture with a preview paragraph and "To continue reading: Subscribe Now or Sign In".
The article says that the boy liked the bionic pancreas so much that he ran away from the investigators conducting the test, and it took them over an hour to get the device back.
Given how much this technology costs it is rather surprising that they have not yet built some form of locator tech into it.
At least, not yet.
I read a blog recently by a woman who took part in the trials. She said she cried when she had to give it back. Apparently that was a very common reaction by participants.
They factored the CGM lag into the equations, niccolo. You gotta look at it, its amazing! I didn't think it would work either, due to how hard it is to take measurements on all the variables involved. They have been trying to build a solid mathematical model, for years, and it just never pans out. They got it this time, though, I think. The math looks solid. You gotta check it out. Fast forward to min 50 to see its predictive capabilities - 25 min. https://www.youtube.com/watch?v=rrtPn6GKMlU
They use a set of really chaotic data, and their predictive model fits it like a kid glove, 25 min out. Thats real tough to do, but people build similar models for flying aircraft, so its conceivable that they did it. If they publish the model, and I'm sure they will, then anybody could implement it.
Fascinating. And I suppose even when there are occasional outliers, which by definition will happen, the AP should be able to correct them far more quickly.
[I'm a subscriber but have no idea how to read it or if I can share it online...]I was encouraged by Ed Damiano stating "the bionic pancreas has to be covered, or it's not going to work" referencing insurance. Although I'm sure the other side is lining up reasons not to cover it as we speak...
The Time Magazine article is simply an update of an ongoing story that dates back a few years and was featured on NPR, USA Today, Diabetes Mine, and many others. For those who are not subscribers to Time, you can read about the device at:
NPR: Father Devises A ‘Bionic Pancreas’ To Help Son With Diabetes
http://www.npr.org/blogs/health/2014/06/16/320309856/father-devises-a-bionic-pancreas-to-help-son-with-diabetes
NPR Audio (MP3) Version:
http://pd.npr.org/anon.npr-mp3/npr/me/2014/06/20140616_me_father_devises_a_bionic_pancreas_to_help_son_with_diabetes.mp3?dl=1
USA Today: Bionic pancreas could help fight diabetes
http://www.usatoday.com/story/news/nation/2014/02/02/bionic-pancreas-could-help-fight-diabetes/5131531/
NBC News: ‘Bionic Pancreas’ Astonishes Diabetes Researchers
http://www.nbcnews.com/health/diabetes/bionic-pancreas-astonishes-diabetes-researchers-n130956
You saying you occasionally get strange high or low readings outta the Dex? I don't think I've gotten those, but maybe I just haven't noticed. Points that are separated from the neighboring points? Maybe it makes you calibrate the Dex during rapid increase/decrease or calibrate more often. That's what I would do. I honestly don't know how they deal with sensor inaccuracy. I would think that's a problem.
If they publish the model, anybody and everybody should be able to implement it. Freeware shouldn't be a problem, unless there's hardware involved, but the real advancement is in the math and I think they will publish that because its coming out of academia. A lot of the code is published already. They are getting funding by the NIH, so we have already kinda paid for it upfront. I'm gonna try to implement the model.
Although the news about this bionic pancreas has been out there for a while, I was really struck by the comments of the participants in the clinical trials. That's what made this article cause me to sit up and take notice.
I agree Melitta, it must be pretty good if people ran away with it or cried when they had to give it back! I mean pretty good in like you feel like you don't have D anymore.
So, there’s like a million people doing this kinda research and they have been for a while. I get all of them confused. Is the ‘bionic pancreas’ the one that’s going through the FDA right now? Or is it a different one?
Acid rock was right, thats the one with glucagon. Is that the one going through FDA? There will be hardware requirements in addition to what we already own. Thats not good. Who sent it to FDA? Anybody got the control algorithm? No can find.
There are a bunch of them working in parallel. I think the Damiano one gets a bit more press, maybe the personal connection but he seems to be moving right along. I think the process is that they have to run them through self-funded trials and then submit the trial results to the FDA to secure permission to sell the gizmo. It seems like the FDA has made progress but, at the same time, things like Nightscout, CGM in the cloud show what people can do if left to their own…devices…
OK, we got alternatives, then. The hang up with new hardware is that it has to go through FDA. Software has no such requirement. If you want the glucagon, then you gotta wait. But, if you can tolerate eating for lows or bumping down basal, then it will be cheaper. I don't think nightscout implements the statistical model. It basically just a Dexcom, right? Me want predictive model.
Software does indeed have that requirement. Dex had to wait for approval from FDA before releasing the 505 update for the Dex G4 a couple months ago. :)
Slap me when I get long winded, but let me be more specific. I can feel myself ramping up for a couple paragraphs, so I apologize in advance.
Dexcom is an ethical hardware company and I want to applaud them. They built the best sensor on the market. What was distinctive, and insightful, about how they did it, as opposed to Medtronic, was that they allowed us access to our own medical data. They know they aren’t the ones that are gonna knock down the dominos, but they set up the dominos so that they COULD be knocked down.
Medtronic wrapped my blood sugar data in a proprietary software package that made it useless. No one, not even I, had access to my data. Only Medtronic. They actively put up a barrier so that their own device was worth nothing. If there is a just God out there, she has secured a place for Medtronic in medical manufacturer hell.
Dexcom’s hugest advancement, wasn’t even the hardware, it was that they had ethics. They allowed me to pull my own data and import it into statistical software for analysis. Now, the entire world can take a crack at it, even me. They knew they didn’t have the solution, so they stood back and let technology move forward, unencumbered.
Dexcom knows that what they built was only one half of the puzzle. The other half is the mathematical model. Now, I don’t think we know who, specifically, is gonna knock those dominos down, but I think we know, for certain, that it’s coming. There have been publications. We’ll have to wait and see if the new models hold up, but I have a strong feeling it will, pending peer review.
Anybody can build any kind of medical app that they want and release it onto the internet, legally. Statistical analysis is not against the law, nor should it be. Thats the other half of the puzzle. What’s great about this is that there is suddenly ample opportunity for the whole world to try and build a valid mathematical model. Math has been holding up development for a long time, as usual. Math is the hard part. Hardware and software are the easy part. Any joker can write the software once somebody figures out the mathematical model. Its coming, boys. I believe the gentleman in IL has got it now. The dominos are starting to fall.