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

20th
Sep 2017
Martin Haeberli
@mhaeberli
Sep 20 2017 00:04
Can apps get real-time estimated glucose values (EGVs) with the Dexcom API?
Currently, the Dexcom API offers EGV data on a three-hour delay. The Dexcom API is intended for software developers to retrieve retrospective CGM data and CGM-derived metrics for use in software applications. Real-time treatment decisions should not be made based on the data displayed to the end user, and the delay helps clarify this. Dexcom settled on the three-hour delay after consultation with FDA.
probably the wrong forum for this. I have already emailed Dexcom. Delay is average 3.5 hours because … But if you USB tether to Dexcom Clarity then you can get real-time. We are here, with openaps, of course making treatment decisions based on real-time EGVs via Share most frequently. But I would also like real-time calibrations - because then can auto-remind to calibrate when needed...
Scott Leibrand
@scottleibrand
Sep 20 2017 00:07

what page did you see that on? I missed it when reviewing their API docs.

perhaps https://gitter.im/nightscout/decoding-dexcom would be a good place to discuss further?

Dana Lewis
@danamlewis
Sep 20 2017 00:14
FYI - I just merged 0.6.0-dev into the dev branch. Which means dev branch is packed with new things and HIGHLY experimental, so huge caution urged for testing this branch. At this point, I wouldn't test without looking at the code diff. There's a few notes in the PR (https://github.com/openaps/oref0/pull/650#issue-256564183), but we need some more help with other documentation and flagging other major changes. So, please proceed testing dev with caution, ask questions, and comment on the PR with notes and tips for others as you move forward with testing.
big thanks to @tim2000s @scottleibrand @sulka (for exponential curves work!) & others for all the behind the scenes work on this branch.
garykidd
@garykidd
Sep 20 2017 00:43
SO minor in the grand scheme but was wondering if its worth scribing an issue relating to inability to read SMBs in NS?
tepidjuice
@tepidjuice
Sep 20 2017 02:39
Hi. I use xdripAPI and a bluetooth tether to my phone to get my cgm data onto my rig. The cgm data is not being sent and in xdrip+ under restAPI it has 'marking records failed'. Same issue that I had a couple of months ago for which I ended up having to reflash my rig.
Scott Leibrand
@scottleibrand
Sep 20 2017 02:40
@garykidd if you have any ideas for better ways to visualize SMBs, it might be worth discussing...
Tim Street
@tim2000s
Sep 20 2017 05:40
@tepidjuice - I’ve a new rig setup - let me test it and see. Fwiw when I updated to the latest Code, xdripaps continued to work on my older rigs.
tepidjuice
@tepidjuice
Sep 20 2017 05:43
@tim2000s thanks I really have to get a back up rig running. I've ordered the parts can have a test rig.
Nalla306
@Nalla306
Sep 20 2017 07:33
been trying for hours with no luck now. Because it happens on both rigs is it likely to be a pump problem? It did work showing pump status in nightscout last night for about 30 minutes.
But now i just get the following errors
.mmtune: mmeowlink.vendors.serial_interface.AlreadyInUseException: /dev/spidev5.1 already in use
CommsException("Could not get subg_rfspy state or version. Have you got the right port/device and radio_type?"
tepidjuice
@tepidjuice
Sep 20 2017 10:22
@tim2000s if it helps debugging... I think it occurs after updating my rig.
garykidd
@garykidd
Sep 20 2017 11:53
Is dexcom server down?
Brian Rabinovitz
@bdr1177
Sep 20 2017 13:16
Capture.JPG
Temp target question: in the graph above, temp target is 80 and the eventual bg is predicted to be 141. The temp basal was just set higher at 0.85, but this still won't be enough to reach a temp target of 80. In fact, the next OpenAPS pill after this shows that 0.95 is required, but it decides that 0.85 is close enough for now so it doesn't change the temp basal. I'm fine with all this, but I'm trying to understand the logic. If OpenAPS is trying to reach the temp target of 80 and eventual bg is predicted to be much higher than that, why isn't it temping much higher (our maxsafebasal would be 1.05).
Chris Oattes
@cjo20
Sep 20 2017 13:19
If it predicts that it'll end up as 141, how much extra insulin do you think is needed to hit 80?
(I can't remember exactly which numbers it uses, but it's likely it's actaully using 134 in the calculation)
Brian Rabinovitz
@bdr1177
Sep 20 2017 13:23
With 0.39 on board and ISF of 175, that would theoretically bring bg down to about 100. So I guess another 0.1 would bring it down to about 80.
Chris Oattes
@cjo20
Sep 20 2017 13:23
134 - 80 = 54 mg/dl. With an ISF of 175, that's a requirement of ~0.308u insulin. To deliver that over 30 minutes, that's an increased in basal rate of 0.616u/hr. Baseline looks like 0.25u, so 0.25 + 0.616 = 0.866u/hr.
Brian Rabinovitz
@bdr1177
Sep 20 2017 13:25
That makes sense, but why doesn't it just deliver that through a higher basal rate over a shorter time?
Chris Oattes
@cjo20
Sep 20 2017 13:27
You can only set basal rates in 30-minute increments. If it sets the basal rate higher, it has to manually cancel the temp basal when the correct amount has been delivered. If you walk out of range of the rig, or its battery dies, or there's interference, or ...., then the pump will end up delivering more insulin than it needs to get to the target range, which may cause a hypo
Brian Rabinovitz
@bdr1177
Sep 20 2017 13:28
@cjo20 Yes, thank you. That makes perfect sense now.
Chris Oattes
@cjo20
Sep 20 2017 13:28
So when it sets basal rates, it only sets a rate that is safe to leave to run to compeletion
Brian Rabinovitz
@bdr1177
Sep 20 2017 13:28
So the limitation is the minimum length of the temp basal
Chris Oattes
@cjo20
Sep 20 2017 13:29
I believe that is what led to the design decision, yes. As the minimum length of temp basal decreases, it approaches SMB.
jaylagorio
@jaylagorio
Sep 20 2017 13:30
Just upgraded to 0.6.0 on my dev rig to test today, already like the look of some of the new settings and the log output
Brian Rabinovitz
@bdr1177
Sep 20 2017 13:31
We don't use SMB because our lowest basal is too low (0.25) so I'm not very familiar with it. Does it work to correct high non-meal bg, or does it only come into play at meal time?
Chris Oattes
@cjo20
Sep 20 2017 13:31
I don't use SMB myself yet
jaylagorio
@jaylagorio
Sep 20 2017 13:32
I use SMB primarily at meal time or on the (lately) rare occasion that I want it to aggressively correct a high
So you could use it either way
Brian Rabinovitz
@bdr1177
Sep 20 2017 13:33
@jaylagorio When you say you use it at certain times, does that mean you need to actively engage it? I thought it worked in the background similar to the normal OpenAPS temp basal adjustments.
jaylagorio
@jaylagorio
Sep 20 2017 13:35
You have to enable it since it's off by default, and you choose any of three scenarios when it kicks in
When you manually bolus, when you have carbs on board, or when you set a temp target
Brian Rabinovitz
@bdr1177
Sep 20 2017 13:35
Got it. Thanks.
jaylagorio
@jaylagorio
Sep 20 2017 13:35
I chose COB and temp targets so if I want to manually bolus to do a correction myself I only have to worry about the normal OpenAPS temp adjustments
Meghan Rutledge
@mar5477
Sep 20 2017 15:29
image.png
I assume I need to delete the corrupt file, but I'm not sure how to get there.
jaylagorio
@jaylagorio
Sep 20 2017 15:39
Can someone describe the predCIs/remainingCIs entries in the logs during SMB loops?
ReadTheDocs doesn't seem to have a description
Chris Oattes
@cjo20
Sep 20 2017 15:46
jaylagorio: I don't know, I've not looked at it, but I would guess it's to do with carbs it thinks you're yet to absorb, and their impact on blood sugar
jaylagorio
@jaylagorio
Sep 20 2017 15:48
Sure, but lack of units is throwing me off. 0.1 carbs remaining to be absorbed? 0.1 units of insulin needed to counteract absorption? That's why I'm looking for more detail.
Chris Oattes
@cjo20
Sep 20 2017 15:50
can you paste an example?
jaylagorio
@jaylagorio
Sep 20 2017 15:50
Sure!
predCIs: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0                       
remainingCIs: 0 0 0 0 0 0 0 0 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0
.1 0.1 0.1 0.1 0.1 0.1 0.1 0 0 0 0 0 0 0 0 0
Chris Oattes
@cjo20
Sep 20 2017 15:52
hmm. Not sure. You could probably look at the code and work it out
jaylagorio
@jaylagorio
Sep 20 2017 15:53
You're right, and I tend to do that - but I ask here first to make sure it isn't an FAQ or the answer isn't apparent first
Meghan Rutledge
@mar5477
Sep 20 2017 16:36
update on corrupt git object: it was under myopenaps directory and there were a lot more corrupt files. Changed the name of the directory and am trying the setup script again
garykidd
@garykidd
Sep 20 2017 16:45
@scottleibrand regarding the less than ideal viewability of SMB in NS...the most simple quick fix that I can visualize would be to simply limit the viewed time frame to a smaller duration--most-likely 60 mins.
katie disimone
@Kdisimone
Sep 20 2017 16:46
@mar5477 that git directory will automatically be rebuilt by the system. You don't need to reinstall.
garykidd
@garykidd
Sep 20 2017 16:47
I think that'd create enough space to better see amounts. Of course something more substantial in work could involve plotting SMBs on the basal line---really they don't need to be shown on the commensurate BG level line
Meghan Rutledge
@mar5477
Sep 20 2017 16:53
@Kdisimone I ran git fsck until I it didn't come back with anymore corrupt objects. But then I was getting "error: HEAD: invalid sha1 pointer" and wasn't sure what to do from there. Googling led me to believe I needed to point back to an okay object, but git pull/fetch/checkout didn't work
@Kdisimone should I just have let time fix things after removing the corrupt objects?
myopenaps# git status came back with "fatal: bad object HEAD"
katie disimone
@Kdisimone
Sep 20 2017 16:56
It depends on where your git corruption was. If it was in myopenaps, I typically do cd ~/myopenaps && rm -rf .git That will remove the directory and it will be rebuilt by the cron. If that (the rebuild) doesn’t happen automatically, then I would remove the openaps directory too and run the script fresh cd && ~/src/oref0/bin/oref0-setup.sh
cd && rm -rf myopenaps to remove to the myopenaps directory
Scott Leibrand
@scottleibrand
Sep 20 2017 16:59
@jaylagorio those are debugging lines I intend to remove before release rather than documenting. they display the predicted carb impact for each 5m period out to 4h. the units are mg/dL per 5m.
jaylagorio
@jaylagorio
Sep 20 2017 17:00
@scottleibrand Unless you plan to have this info in some other format, I think if you wanted to leave them in but add units that'd be pretty cool!
Scott Leibrand
@scottleibrand
Sep 20 2017 17:00
they give me a way to visualize how much of the carb impact is coming from actual observed carb absorption projected out in a \ shaped predCI curve (the first line) vs. remainingCI that isn't projected to absorb over ~4h and is being projected as part of the /\ shaped remainingCI curve
Meghan Rutledge
@mar5477
Sep 20 2017 17:01
rm -rf .git looks like it worked? in the meantime, changing the directory name, and running the setup script again allowed the loop to start work again
jaylagorio
@jaylagorio
Sep 20 2017 17:01
I like it!
Scott Leibrand
@scottleibrand
Sep 20 2017 17:01
@jaylagorio I think for normal users, looking at the COB purple line prediction in NS gets them the info they need. I think this level of detailed visualization will only be needed as long as I'm playing around with this part of the code
if I decide I no longer need it, I'll remove it. If I fall in love with it and think it's worth the extra space in the pump-loop.log, I'll make it clearer what it's showing. :)
jaylagorio
@jaylagorio
Sep 20 2017 17:02
Haha, I think that's a good compromise
Speaking of visualization, I think I'm seeing some lag for what the loop does vs. when it gets to Nightscout. As an example:
  1. Do the loop, SMB delivers 0.2 U
  2. Loop finishes. Nothing appears on Nightscout. Refresh Nightscout just to make sure - no change.
  3. Next time through the loop, Nightscout gets updated.
I've noticed this for basal changes too, the run through the loop when it happens doesn't deliver the change to Nightscout, only the run through after it already happens
katie disimone
@Kdisimone
Sep 20 2017 17:04
@jaylagorio That’s typical NS behavior for DIY looping…a lag from delivery of info to confirm and recording, finally visualization of the info. Happens in Loop environment too.
jaylagorio
@jaylagorio
Sep 20 2017 17:05
Yeah, I don't think it's brand new, just more noticeable now that I'm using SMB and trying to eyeball what the IOB is based on Nightscout pills
Is that something that can be fixed?
katie disimone
@Kdisimone
Sep 20 2017 17:06
@jaylagorio not so far as I can see
Scott Leibrand
@scottleibrand
Sep 20 2017 17:07
we would have to add another pumphistory refresh to the end of oref0-pump-loop
I'd avoided doing that until now, but if an SMB was delivered, there's not much else that needs to be done right away, so maybe it'd be a good idea. will have a slight battery hit, but maybe worth it
jaylagorio
@jaylagorio
Sep 20 2017 17:10
Can you do a shorter pumphistory refresh when you know you just did a change? Or format the SMB/temp change/etc as a treatment when it happens and deliver that to Nightscout?
Scott Leibrand
@scottleibrand
Sep 20 2017 17:10
we already do a short pumphistory refresh now. no reason to re-do the pumphistory-24h refresh
I don't want to assume something got delivered if it might not have: that's why we wait until it's confirmed present in pumphistory before uploading
one other thing is I'd need to wait until the SMB has finished delivering before refreshing pumphistory. need to figure out how to know how long that is...
maybe I'll just check if it's still bolusing, and if so wait 30s before retrying once: smb_verify_status || ( sleep 30 && smb_verify_status ) ) && gather
jaylagorio
@jaylagorio
Sep 20 2017 17:16
Yeah, agree on not doing the 24 hour history. I think it was written that oref1 does a check of the reservoir, does the SMB, then when the SMB finishes confirms by checking the reservoir again. Would that confirmation not be enough?
But you're totally right that some form of confirmation should be done to make sure what gets pushed to Nightscout actually happened
I didn't know whether temps are double-checked as to whether they were successfully delivered or not. Maybe if the temp is sent and that action doesn't error out that is indicative that it was delivered successfully?
Scott Leibrand
@scottleibrand
Sep 20 2017 17:17
too many additional points of failure if we start guessing about pumphistory contents. would be a pain for deduplicating too. better to just refresh it and see if the bolus/temp is actually there
jaylagorio
@jaylagorio
Sep 20 2017 17:19
That sounds best. If you do a short pumphistory refresh I wouldn't worry too much about the battery vs. the load it already puts on the pump and I think the immediate feedback is important with SMB.
I could be comfortable skipping that with temps
Scott Leibrand
@scottleibrand
Sep 20 2017 17:20
:+1:
Tim Street
@tim2000s
Sep 20 2017 17:39
I guess the pump wearer gets instant feedback on SMBs as they feel/hear the pump. Maybe not amount, but if I need to worry about that, that’s what pushover is for.
scottleibrand @scottleibrand doesn't wear the pump, so I've already implemented this in a branch for my own benefit. :)
Dana Lewis
@danamlewis
Sep 20 2017 17:47
Lol
Parents will probably like this too
garykidd
@garykidd
Sep 20 2017 18:09
I've gotten plenty of SMBs without noticing visual or vibration feedback=good thing for me though.
thisistheusername
@thisistheusername
Sep 20 2017 18:30
@danamlewis @scottleibrand @Kdisimone How many people have been able to loop with a 512/712? I am at a deadend with my 712. If the number of 512/712 success stories is negligible, would it be better to drop those pumps from the supported pumps list?
Scott Leibrand
@scottleibrand
Sep 20 2017 18:32
sounds like we need a volunteer who could work on getting a 512/712 set up with a dev rig, and make the necessary changes to oref0-setup etc. to support it more automatically
Dana Lewis
@danamlewis
Sep 20 2017 18:33
There are half a dozen or so in past two months. I think the challenge is people scavenge and find other pumps that they switch over to, keeping their 12 for backups. Other challenge is not having a volunteer to help test and doc things on 12s.
katie disimone
@Kdisimone
Sep 20 2017 18:55
@thisistheusername the last I saw, your files looked significantly different than the files that the docs spell out. How closely did you use the docs to set up your 512 and rig? As a troubleshooter, it is hard to know where issues may be when people use a method so different than the recommended way. I'd be happy to take a look at things again, have you posted new versions of the files? Did you successfully edit your openaps aliases? Did you set your remote devices to on in the pump and enter a dummy remote ID?
thisistheusername
@thisistheusername
Sep 20 2017 19:34
Where is the documentation on the remote id? I asked about that initially but was told it didn't matter.
garykidd
@garykidd
Sep 20 2017 19:57
wondering if it's normal/expected to see increased basals before seeing appearance of SMBs when following normal meal pattern of bolusing a buit upfront...and inverse as it unwinds ..meaning if you had some SMBs as carbs continue to metabolize as BG drops....first you'd see SMBs and then temp basals increase as you encroach on target BGs?
Scott Leibrand
@scottleibrand
Sep 20 2017 20:24
I'm not following. If you don't have a lot of IOB yet, oref0-pump-loop first sets a high temp, and then tries to SMB, and then lowers the temp (eventually down to zero) on subsequent runs once there's enough IOB from the SMBs.
garykidd
@garykidd
Sep 20 2017 20:25
right--you are actually following....
thisistheusername
@thisistheusername
Sep 20 2017 20:33
@Kdisimone
@Kdisimone could you point me at the documentation regarding the remote settings for the x12 pumps?
katie disimone
@Kdisimone
Sep 20 2017 20:36
@thisistheusername I don’t have one, and yes, people have said that is not a required step. I tend to do the things people suggest, even when they are “optional” when troubleshooting. I asked you several follow up questions. Can you answer them too?
Dana Lewis
@danamlewis
Sep 20 2017 21:32
@thisistheusername also, can you confirm if you were trying to run this with SMB, or just a regular initial setup?
I am testing the 512 instructions, and just confirmed that x12 won't be compatible with SMBs
but fine for regular looping
Dana Lewis
@danamlewis
Sep 20 2017 21:50
finding some other interesting x12 potential issues, too.
Scott Leibrand
@scottleibrand
Sep 20 2017 21:57
looks like x12 looping was broken with the switch to oref0-pump-loop in 0.5.3, but we didn't have any x12 testers to notice that.
should be fixable though
Dana Lewis
@danamlewis
Sep 20 2017 21:57
yep, working on it..
daveklingler
@daveklingler
Sep 20 2017 22:00
Okay, so I'm trying to get my personal hotspot working again on my iPhone. Looking at the oref0-online.sh script, I noticed that it wasn't finding an IP address for bnep0:
root@binky:~# ifconfig
bnep0     Link encap:Ethernet  HWaddr 98:4f:ee:04:1c:ca  
          inet6 addr: fe80::9a4f:eeff:fe04:1cca/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:1 errors:0 dropped:0 overruns:0 frame:0
          TX packets:6 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:67 (67.0 B)  TX bytes:276 (276.0 B)

lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:65536  Metric:1
          RX packets:85 errors:0 dropped:0 overruns:0 frame:0
          TX packets:85 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:12012 (11.7 KiB)  TX bytes:12012 (11.7 KiB)

usb0      Link encap:Ethernet  HWaddr 02:00:86:27:c3:e5  
          inet addr:10.11.12.13  Bcast:10.11.12.255  Mask:255.255.255.0
          UP BROADCAST MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:0 (0.0 B)  TX bytes:0 (0.0 B)

wlan0     Link encap:Ethernet  HWaddr 78:4b:87:ac:5a:e8  
          inet addr:192.168.1.12  Bcast:192.168.1.255  Mask:255.255.255.0
          inet6 addr: fe80::7a4b:87ff:feac:5ae8/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:409243 errors:0 dropped:0 overruns:0 frame:0
          TX packets:419070 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:128221700 (122.2 MiB)  TX bytes:53393204 (50.9 MiB)

root@binky:~# ip -4 -o addr show dev bnep0 | awk '{split($4,a,"/");print a[1]}'
root@binky:~#
Scanning over the ifconfig output, I realized I didn't have an ip4 address, and I didn't know how to get one. @scottleibrand , can you tell me what I'm missing, please?
That is, I'm not picking up an ip4 address for BT.
Savek-CC
@savek-cc
Sep 20 2017 22:04
@danamlewis can you extract the number of loopers in .de from your dataset? (either total or even better by type of loop?) I'm going to give a presentation next month and would like to include this information...
Dana Lewis
@danamlewis
Sep 20 2017 22:04
always kept updated here: https://openaps.org/outcomes/
daveklingler
@daveklingler
Sep 20 2017 22:05
Or anybody else who can help. :)
Dana Lewis
@danamlewis
Sep 20 2017 22:05
do you mean in germany specifically? @savek-cc
Savek-CC
@savek-cc
Sep 20 2017 22:05
exactly @danamlewis - so the part of (n=1)*401+ ;)
Scott Leibrand
@scottleibrand
Sep 20 2017 22:06
@daveklingler you can try manually running the dhclient that oref0-online does to get an IPv4 address via DHCP
Dana Lewis
@danamlewis
Sep 20 2017 22:06
I can skim the list and give you an eyeball account. no good automated way to do it since everyone fills it out so differently. do you want that now, or closer to your presentation?
won't be by loop-type, will just be eyeball count
Savek-CC
@savek-cc
Sep 20 2017 22:06
guess once the ac combo will be supported by aaps it might increase in .de
eyeball right now is fine...
or any time within the next 2 weeks or so - when you have the time
daveklingler
@daveklingler
Sep 20 2017 22:08
@scottleibrand , ah, thanks, that's what's not happening. I ran dhclient and it pulled an IP. Now I just have to figure out why it's not happening in the script.
Dana Lewis
@danamlewis
Sep 20 2017 22:09
@savek-cc ~34
Savek-CC
@savek-cc
Sep 20 2017 22:09
thanks @danamlewis
Dana Lewis
@danamlewis
Sep 20 2017 23:13
boo yea! temp set on the 512. and bugs fixed for 512. @thisistheusername if you want to go pull a fresh oref0 master, I bet it'll work now.
(cut the 0.5.4 release to include the x12 fixes)
these instructions do work well still: http://openaps.readthedocs.io/en/latest/docs/Build Your Rig/x12-users.html
(same URL with spaces encoded as %20 so it's clickable)
jaylagorio
@jaylagorio
Sep 20 2017 23:45
Howcome the x12s don't work with SMB?
Dana Lewis
@danamlewis
Sep 20 2017 23:46
the status checks we rely on don't have the same format status files in the x12
so if someone wants SMB to work with x12, would need to do the work to help figure out suitable alternative for safety checks
jaylagorio
@jaylagorio
Sep 20 2017 23:46
That's a shame - do you think that's something that can be overcome or is there data that's straight-up missing?
Dana Lewis
@danamlewis
Sep 20 2017 23:46
the suspend/bolusing checks we use for x15/22/23/54 fails on 12s because it has a different status.json
jaylagorio
@jaylagorio
Sep 20 2017 23:47
Gotcha
Dana Lewis
@danamlewis
Sep 20 2017 23:47
honestly, didn't dig in it too hard - was focusing on a) fixing basics for x12s (done now) and now b) trying to automate the set up process for x12
jaylagorio
@jaylagorio
Sep 20 2017 23:47
Yeah, I think those are higher priorities for sure
Dana Lewis
@danamlewis
Sep 20 2017 23:47
so if you take alook at what we just changed in 0.5.4, that'll show you where to maybe start digging in if you want to tackle it
I just want to keep x12 in the family for basic looping
jaylagorio
@jaylagorio
Sep 20 2017 23:48
Unfortunately I don't have a x12 to bang on
But today with 0.6.0 was actually really good. No meal boluses and I didn't go over 200
Dana Lewis
@danamlewis
Sep 20 2017 23:48
nice :)
jaylagorio
@jaylagorio
Sep 20 2017 23:48
image.png
image.png
^^ Graphs with a few annotations
Scott Leibrand
@scottleibrand
Sep 20 2017 23:48
w/ fiasp or humalog/novolog?
Dana Lewis
@danamlewis
Sep 20 2017 23:48
fiasp? or just regular insulin?
jaylagorio
@jaylagorio
Sep 20 2017 23:48
You guys weren't kidding about the sharp rises and sharp drops now
Novolog U100, regular
Dana Lewis
@danamlewis
Sep 20 2017 23:49
ya
oh interesting
jaylagorio
@jaylagorio
Sep 20 2017 23:49
Is Fiasp available in the US?
Dana Lewis
@danamlewis
Sep 20 2017 23:49
yea on fiasp it's even sharper in both directions. the first few times I overcorrected w/ it on both sides
no
you can order online from pharmacies in canada, though
jaylagorio
@jaylagorio
Sep 20 2017 23:49
Well, "regular" meaning whatever I've been using in my pump for 8 years. I think there's actually an insulin type called "regular" and I don't know if this is it
Dana Lewis
@danamlewis
Sep 20 2017 23:49
ha, no
jaylagorio
@jaylagorio
Sep 20 2017 23:49
It's an orange box and I know it's called Novolog U100
Dana Lewis
@danamlewis
Sep 20 2017 23:49
"the usual ole Humalog/Novolog" is what I usually mean by regular
yep
jaylagorio
@jaylagorio
Sep 20 2017 23:50
Yeah, that's the one
So I don't have access to Fiasp but I'd be interested in trying it with this
Scott Leibrand
@scottleibrand
Sep 20 2017 23:50
Apidra is also pretty similar. in all three cases we recommend the "rapid-acting" curve
Dana Lewis
@danamlewis
Sep 20 2017 23:50
it's valuable to see 0.6.0 tested with Novo/Humalog, though, so will be curious to see more feedback from it
jaylagorio
@jaylagorio
Sep 20 2017 23:50
I noticed that the first thing I did today was a meal temp for a bit and that rise was less sharp than the others during today, which I didn't do a meal temp for
Dana Lewis
@danamlewis
Sep 20 2017 23:50
did you switch to exp curves, or still bilinear?
jaylagorio
@jaylagorio
Sep 20 2017 23:51
But I was pretty aggressive in fighting things down so maybe that's where some of the steep drops came from
Scott Leibrand
@scottleibrand
Sep 20 2017 23:51
https://rxcanada4less.com/ will ship Fiasp to the US
jaylagorio
@jaylagorio
Sep 20 2017 23:51
It's still bilinear, I didn't immediately see where to switch it and I figured since it's the same insulin I shouldn't change that
Scott Leibrand
@scottleibrand
Sep 20 2017 23:52
I would recommend switching to exponential curves before relying on no-bolus eSMB
jaylagorio
@jaylagorio
Sep 20 2017 23:52
Interesting link! That's a little outside my price range but it's good to know it's possible. I would consider switching after seeing some more info online and bringing it up with my doctor
So there's a knowledge gap for me - what's the difference if it's the same insulin?
Scott Leibrand
@scottleibrand
Sep 20 2017 23:53
there are cheaper options if you (or your mail forwarding service) are in Canada
jaylagorio
@jaylagorio
Sep 20 2017 23:54
Just looked up my current type there, they want more for one box than my insurance charges for three boxes, so I think I'm going to stay put for the time being
Scott Leibrand
@scottleibrand
Sep 20 2017 23:54
yeah, if you don't have insurance canada is cheaper, but usually more than copays
jaylagorio
@jaylagorio
Sep 20 2017 23:55
It's something I've always kept in the back of my head in case I lose my insurance but I'm glad to have a link people find reputable
I saw the thing about the different curves in the 0.6.0 Issue Dana posted last night but what's the difference between the curve types if the insulin doesn't change? And how do I go about setting a different one?
jaylagorio
@jaylagorio
Sep 20 2017 23:58

So I'd use

"curve": "rapid-acting"

For Novolog, or is that for Fiasp?