These are chat archives for nightscout/intend-to-bolus

24th
Mar 2015
diabeticgonewild
@diabeticgonewild
Mar 24 2015 05:45

Too sick to sleep. Just not feeling well in general. :( I had to max out my dose of medication, which is sort of concerning, even though it has "worked" so far.

Going to try to work on AP stuff, instead of school work! Haha!

But, I have worked way too hard on my engineering degree to take my education for granted or to just give up on it, so it's a hard choice for me to make, even in the short-term. As for staying motivated to get my degree in engineering despite all of the terrible medical stuff I have going on, my parents, some engineering friends, and a couple of mentors have been my strongest influences, but my parents would get the most credit by far... I should thank my mom more often. I can't thank my (late) dad, but I am sure he would be very happy with me.

Sorry for rambling, but it's good to be grateful sometimes...

diabeticgonewild
@diabeticgonewild
Mar 24 2015 06:37
OK, I just found something awesome! So you all know how a limitation of the paper (Stochastic, Haidar, Hovorka, et. al.) I am using is the lack of plasma insulin concentrations which can lead to potential (slower or a low possibly no) convergence for the stuff I am programming. Well, in a new paper by Hovorka, they are using this document to calculate plasma insulin concentrations now, because they could not get them for lispro/Humalog. The best part, is the paper it comes from, is for type 2s, which would have more variables, and they tend to take more insulin anyways, so it is a good sign.
So, I should be set! Pharmacokinetics of insulin lispro in type 2 diabetes during closed-loop insulin delivery
also, all of my questions have been answered through reading and research...I just need to email the author a few questions.
About approach and stuff.
diabeticgonewild
@diabeticgonewild
Mar 24 2015 06:45
In the link I provided, look at the equations under insulin measurement in heparinized plasma and serum , two equations and 2 unknowns = solvable. Also, mIU is milli International Units, and insulin I know is measured in some variant of international units, for what we call "units" in regular speak.
diabeticgonewild
@diabeticgonewild
Mar 24 2015 06:53
This paper from Hovorka was published in November...this totally makes my month!
diabeticgonewild
@diabeticgonewild
Mar 24 2015 10:27
Almost had a heart attack...thought I didn't get a 100 on my homework assignment that I had to submit...but I did get a 100. I love my university.
diabeticgonewild
@diabeticgonewild
Mar 24 2015 12:19
Sooo sick this morning...just took a cocktail of medicine in hopes of being able to sleep some of being sick off
Toby Canning
@TC2013
Mar 24 2015 13:06
I think that may be the wrong link above?
Chris Oattes
@cjo20
Mar 24 2015 13:58
TC2013: you sat there with your ear to the pump listening for the small clicks for 8 hours?
diabeticgonewild
@diabeticgonewild
Mar 24 2015 13:59
Maybe I will check the link in a little bit
Toby Canning
@TC2013
Mar 24 2015 14:19
@cjo20 Well, along with some more exciting stuff, like using a stopwatch app with a lap timer to time the intervals... LOL
Chris Oattes
@cjo20
Mar 24 2015 14:20
TC2013: that must have been thrilling :P I imagine it was a little like chinese water torture
kenstack
@kenstack
Mar 24 2015 15:32
@tc2013 are you using a 722 or a 5 series ? I'm wondering if anyone has confirmed that 10 strokes per unit is correct for boluses when using mm-bolus.py from carelink. I did some gross testing last year and it seemed correct but I'm getting back into the ap stuff and I was hoping someone did some detailed tests. I was thinking of doing some but to get the accuracy we need I'll need a mass balance.
@bustavo what are you using for strokes per unit ?
kenstack
@kenstack
Mar 24 2015 15:38
@TC2013 I just saw your earlier post and saw you are using a 522 sorry :)
Toby Canning
@TC2013
Mar 24 2015 16:49
@kenstack This is what I learned from my own testing
1 stroke delivers 1/10 of the bolus amount, 2 strokes 2/10, and so on. What I didn't test is if you can deliver less than .1 bolus amounts by changing it to 5 strokes.
timberFrame
@timberFrame
Mar 24 2015 17:32
Hi diabeticgonewild, Reading the section in the paper you linked, I just see a linear conversion between measurements of insulin in serum and in plasma. Are you after a model that predicts blood insulin concentration per time? Like Rodbard, etc? Ultimately, the model will have a fudge factor that relates the concentration of insulin in the blood to insulin action on glucose. This is an insulin sensitivity factor. It varies considerably within the same individual at different times. I can provide some examples.
kenstack
@kenstack
Mar 24 2015 17:36
@TC2013 thanks ! I think the default is set to 15 according to the output of mm-bolus.py - so would it deliver a 1/15 th? Sorry I'm confused :)
Jim Sifferle
@jimsiff
Mar 24 2015 19:28
Is anyone doing any DIY closed loop work with tslim? I just read that they may be rolling out newer software to the pumps at some point. I wonder what their delivery mechanism is... whether the stock firmware can be saved, dissected, and understood... and the bluetooth chip enabled to communicate to a controller device such as an iPhone as done in the Ed D trials.
Toby Canning
@TC2013
Mar 24 2015 21:55
@kenstack hmmmm, my memory may be playing tricks on me, but what I remember is that strokes needing to be 10 for full delivery. About a week ago I broke my second carelink stick and I was waiting to replace it in hopes of getting the cc1111 up and running. So, I can't do any retesting to confirm.
@diabeticgonewild Can I take you up on your offer to lend me a Carelink stick for a bit?
Ben West
@bewest
Mar 24 2015 22:15
@jimsiff if the upgrade is done in a home setting, it can be sniffed/captured/analyzed/reproduced
it's not clear what mechanism they are using to "disable vs enable" the bluetooth
but if it's similar to what medtronic did to their builds, the definitions and everything needed to even recognize the bluetooth from software POV is likely removed from the firmware itself, and that new firmware is provided to define and make use of the hardware
in other words, the most likely theory I know of is that it does require full firmware flash, and might not be able to be "toggled" on
but I haven't got any hard evidence which points either way
kenstack
@kenstack
Mar 24 2015 22:23
@TC2013 thanks !! Will do some large boluses and check it this week - may buy a mass balance to do some more detailed experiments