After one week with #OpenAPS, I can see why everyone that is using it can’t stop talking about it. It requires significantly less attention than traditional diabetes management, but all of the information is available to you all the time. I have had some amazing results so far and I have not enabled any of the advanced features yet.
Here are the stats; the left shows data before the artificial pancreas, the right shows after. I should also mention, again, that the week prior was what I would have considered an ‘outstanding week’ before #OpenAPS:
Here are some of my lessons learned:
- Never rely on internet connectivity or servers: I heavily use the ‘offline’ looping option mentioned here at night and during train rides into the office. I had a few instances where I lost internet connectivity and went high or low because the device couldn’t finish its thought. Which leads to:
- Highs and Lows still happen, they are less severe and mostly treated by leaving it alone, but I think I still need to figure out how to work with the new system and not over-do food or insulin to correct the situation.
That’s it for now, very excited about the results so far and hope to have continued improvement!
I started with OpenAPS Thursday night. I gave the system a bit of a workout over the weekend. Let’s take a look first at what I considered a ‘good week’.
These are the 7 days leading up to the weekend. I had my endo (diabetes doctor) appointment on Tuesday with some adjustments, it turned out to be an outstanding week.
Now, before I get into the numbers for OpenAPS, I have to explain a couple of things. 1) Weekends are realllllly hard for diabetics. Inconsistent meal times, meals out mean higher fat, which, in turn, makes it harder to cover food with the right amount of insulin. In addition to this, there are differences in activity, so that also makes it harder to manage. Anyway, my weekend consisted of Thai food, apple picking, pizza, and Chinese food. These foods are notoriously hard for diabetics to cover properly. Well, take a look:
My percentage in range is higher for my ‘rough weekend’ than what I had thought an ‘outstanding week’ looked like before. I’m going to end up making some additional adjustments to how things are working for me, but this is VERY promising. I look forward to continued success with the system :).
- Lows don’t bottom out like they used to. A very small carb correction gets me back to where I need to be. I’m sure this is due to the fact that basal (base-line insulin) is suspended for low blood sugars
- Highs don’t get as high. For the same reason that lows don’t get as low. Basal insulin changed based on your blood sugar, making boluses pack a bigger punch which leads to:
- Some of my insulin to carb ratios (ratio used to determine how much insulin to give yourself based on food consumed) have likely been serving multiple purposes. Covering both ‘short-term’ and ‘long-term’ carbs. OpenAPS does a reasonable job covering the longer term carbs. I still need to experiment with this a little.
All of the parts have arrived! The only component not pictured here is the case/hard shell that will hold everything together. BTW, not using a 3d-printed case can also drastically reduce the total cost of this rig. I’ve seen others modify a Tic-Tac box to hold everything.
My initial reaction to the materials was that everything was much smaller than I anticipated. Here is the Explorer Board, where everything gets connected (excuse the dirty quarter):
Here is the battery, I thought this would have a bigger footprint (excuse the dirty quarter, again):
Next, I followed the instructions on the Understanding Your Rig page in the OpenAPS documentation. Afterwards, I followed the step by step guide for Installing OpenAPS.
The couple of things missing from the lists are:
- Tweezers (used to assemble board)
- Small Phillips Head Screw Driver (used to assemble board)
- Dexcom Charging Cable (used to connect, power, charge device)
- You should press enter a second time on the first command. It will look like nothing happened after you type the ‘screen’ command, but it did, you just have to press enter.
- In order, you enter your network name/password, then you specify your timezone. It’s a small departure from the documentation. After this, you enter what you want the name of your rig to be (which is omitted in the screenshots in the documentation).
- The timing is a little off too, it says 10+ minutes, plus is right. It took about 25 minutes for me.
Anyway, when I was done, and I had watched the logs for a bit, it looked like this:
And now, I’m going to walk around with this in my pocket:
And now my Nightscout looks like this:
I’ll post an update in a few days to share my experience with using the device.
Yay! My old Medtronic pump is compatible with the OpenAPS framework, now it’s time to shop.
These prices are as of this posting. With Tax and Shipping (Massachusetts) the total is: $268.00!!
You might be saying, “Wow, that’s a lot of money!”. I did add the expedited shipping and processing which wouldn’t be necessary if you are patient, but I’ve had diabetes for 27 years and !
OpenAPS is the Open Source Artificial Pancreas System. It offers an ‘open source’, DIY artificial pancreas using make-it-yourself electronics to grab information from a Continuous Glucose Monitor (CGM), data input (carbs, exercise, really anything that would affect blood sugar), and makes changes to an older model insulin pump over a radio frequency. All-in-all a very cool project and a way that many people have greatly improved their diabetes management. You can check out the hashtag on twitter to see some real-time examples of what people are experiencing with it. This is a good one, as of this posting:
Reading through the OpenAPS documentation is a little daunting, but I came to the conclusion that I could get started by setting up my own Nightscout server. Nightscout (CGM in the Cloud) is an open source, DIY project that allows real time access to a CGM data via personal website, smartwatch viewers, or apps and widgets available for smartphones. The step-by-step instructions were great! The documentation walks through everything you need to do to get started. Because I use the Dexcom G5, I set up the bridge to Nightscout and started seeing my CGM data within a few minutes.
Serendipitous, you can graph the basal rate with your blood glucose levels, making Dexabasal totally unnecessary. The documentation does omit the parameter for the setup. You can add the following to your app variables:
This is what it looks like when configured (get off my back, I’m recovering from a low):
When I get home from my work travel I’ll confirm that I have the correct version of insulin pump and keep everything documented as I continue through this journey!
You can see how I’m doing at ANY TIME by visiting my Nightscout server.
The Dexcom G5 continuous glucose monitor (CGM) is awesome! It allows you to check your iPhone for what is effectively your blood sugar every 5 minutes. It provides the most accurate (for me) readings of any other CGM.
There are other benefits as well. Dexcom Clarity is a reporting and trending analysis tool that provides insights into whether or not I am at risk for hypoglycemia (low blood sugar) as well as whether or not I tend to have higher blood sugars overnight. Overall, this is a great tool.
While these trends work for everyone, I would like to get a better idea of what my basal rate* is along side the trend to ensure that my baseline is appropriate given my day-to-day activities. I’m fortunate that Dexcom has a developer portal, enabling technical folks (like myself) to create apps that are ‘built-for-purpose’. In my case, I’ve started researching visual technologies to combine my basal rate with the Dexcom data! The codename for this project is: dexaBasal. Stay tuned for updates as I progress in creating this app!
* — Basal rate details