The one-year Loop milestone, how did it perform?

Today marks my first anniversary of closing the loop using the do-it-yourself automated insulin dosing system, Loop. This is a hybrid system in that one must announce meals and tell the system how many carbs I’m eating and when. Loop examines the blood glucose level, insulin on board, insulin sensitivity, and carbs on board every five minutes and then decides whether to increase, decrease, or maintain my pump’s basal rate. Here’s my post about Loop from one year ago.

I am thrilled with the results Loop gives me. To compare my pre- and post-Loop performance I am displaying the Dexcom Clarity 90-day standard day report for just prior to starting Loop against the same 90-day period one year later.

Before Loop

As this graph depicts, one year ago I was struggling with a pattern of evening and overnight hyperglycemia. I was actively attempting to make changes to my pump settings, eating, and exercise habits to rein in these hypers.

After Loop

This data set shows improvement in three of my four blood glucose goals: time in range, standard deviation and average glucose. My time hypo went up by 1% in the post-Loop numbers.

During this time my A1c improved from the low 6% range to the mid 5% range, my lowest A1c’s in 33 years of living with diabetes.

Adopting Loop was one of the best diabetes treatment tactics I’ve ever used. It ranks up there with using a CGM, insulin pump, and rapid acting analog insulins. It’s also comparable to changing my way of eating to low-carb, high fat. At this point, I would do everything in my power to stay on this system.

While my results have been measurably better, my overall effort has gone down. Loop’s overnight performance has been dramatic. I often wake up under 100 mg/dL (5.6 mmol/L) without any hypoglycemia.

I am covered by Medicare insurance and I am lucky to also be covered by supplemental insurance provided by my former employer. This insurance is not as good as the lettered (A thru G) supplemental plans except that it has been covering my CGM and supplies. I’m using the Dexcom G4+Share system and I’m hoping that not upgrading to the Dexcom G5 will shield me from Medicare’s rules against using any device other than the Dexcom receiver to display BG results.

Overall I am happy with Loop and my personal BG results would degrade if I had to drop it for any reason. I have purchased some old MiniMed pumps to serve as a backup if my current MM 722 pump fails.

I love looping and I’m hoping that any future replacement I make will be a step up in performance.


Excellent performance! I would love the opportunity to try this. I do fairly well, I know, with a (for lack of a better term) “traditional” pump, but it’s not without significant effort, tweaking and semi-constant attention paid to it. Don’t currently have the resources needed to get things set up. I’m afraid that when the various commercial products become available, they will, like the Medtronics 670 (which I know isn’t quite the same thing) , have preset, unchangeable parameters that are likely to be not to my liking, since I aim for tighter control.


That graph brings a tear to my eye! I can’t believe that can happen. But I have talked with people on the 670g and seen their graphs and realize it can happen. That kind of control is possible. And that kind of control without having to spend every waking moment thinking about and dealing with my blood sugar, insulin levels, exercise times and food oh lovely food! It gives me goosebumps when I think about what I had to do when I was first diagnosed to where we are now.
The big problem now is making this kind of advancements available to everyone who wants it.
I am looking so forward to more of these closed loop pumps hitting the market. My pump warranty is up next month but I will wait and see what happens in the next year or so.
Love you graph! Gives us all hope. Thanks for sharing!
And Happy National Diabetes Day! Hope we are all wearing blue and helping make people more aware!


Thank-you for your comments @Thas and @Sally7. I try to limit these types of performance posts to a reasonable frequency since I realize they not only inspire but also might discourage some others. We all must play our own game and comparing results, especially A1c’s, is not useful.

I remember thinking a year ago today how appropriate it was to close the loop on Banting’s birthday!

What I find interesting is that your CGM graph from looping is really just a flattening of your own performance, so the loop is just doing what you’d be doing if you had the time to nano-manage 24 hours a day. Unfortunately, one could only do this with no breaks for sleep, or bathroom breaks or having any type of life whatsoever.

(I like sleep!)

1 Like

I agree with your analysis. While I likely give more time to my diabetes than the average, I’ve never been capable or willing to devote my attention to it every five minutes. In this regard, Loop’s night-time performance blows me away!

I have determined that I don’t want to loop and although I am relatively tech-savvy, I am not sure that I could figure it out. My old Medtronic pump is happily looping with a Minneapolis friend who as payment helped me install xDrip on my iPhone. I am still not using Dexcom supplies purchased by Medicare and have been playing around with different platforms. Sometimes I think xDrip does better than the Dexcom algorithm but mostly can’t stand seeing two different numbers and trends. My favorite combination is my iPhone and Apple Watch with the G5 app along with my Tandem X2 pump giving numbers but no alerts programmed. Unfortunately both are disallowed with Basic Medicare.

I hope that your insurance, @Terry4, will continue to allow you to get the supplies you need to continue on your looping ways.


Congrats, Terry! So exciting! Loop is really amazing.

Caleb had trouble remembering site changes. He switched back to OmniPod about a month ago. Things have gone better than we expected. We had good habits pre-Loop, although we got a little rusty. It’s easy to get used to not thinking too much about diabetes.

Caleb’s bg results are similar with Loop and without Loop; he’s always had very good control with a 6.0 or below A1c. It takes more thought without Loop. The biggest difference I see is having to be more patient about making basal and IC changes. With Loop, once we dialed in his settings, there was very little need to make these adjustments - Loop compensated. I actually wondered how much Loop masked the need for dosing changes.

We were worried about going back. I’m happy to say, it wasn’t quite as bad to get back on the bike as we thought.


My Vibe is out of warranty and I am approved for a replacement DanaR. Just waiting for the paperwork to go through on the pump order.

At the moment I am running AndroidAPS in open loop mode using the Vibe as a virtual pump. I can already see the differences. Even in open loop I went 7 days without any readings lower than 3.8.

1 Like

That’s impressive! We tried open-looping at first. It was hard to keep up with!

I’ve found that my Loop basal rates just have to be close enough so that the Loop variable basal rates are within reach to redirect an errant BG back into range. I only have three Loop basal rates and they’re all bunched from 0.6-0.8 units/hour. Loop does well adding and subtracting from there.

I also struggle to remember my pump site changes. I log them into Nightscout so I can keep track of the cannula age but that function doesn’t actively remind me to do it. I used to use my calendar function to prompt site changes. I’m thinking that is probably the best way for me to dependably change my pump sites on time. I’m going to get my calendar involved again. Getting the reminder on my cell phone does the trick for me.

1 Like

How often do you make manual basal rate changes running open loop? One or two per hour when you’re awake? That’s impressive that even open loop could reduce hypos so well.


Hey @Terry4, thanks for the update - very glad to see such fantastic results with Loop. After almost 2 years of running DIY closed-loop systems now, I still find myself in disbelief at how much progress has been made by the DIY community, with not one but three closed-loop systems available: OpenAPS, AndroidAPS, and Loop, all way ahead of anything commercially available IMO.


I find it hard to construct a scenario where the commercial systems ever catch up! The gap seems to grow larger every day.

1 Like

Wow- Terry are you saying you never went below 65 or above 130 from August to November?

No. The graph above just shows data from the 15th through the 75th percentile. I definitely ranged outside of those limits, many times! I spent 4% of my time below 65 mg/dL and 14% of my time above 130 mg/dL. The goal with this graph is to squeeze these layers as close together as possible and keeping the outer bands above the hypo line and mostly below the upper limit.

Here’s a current graph that shows the 10th to 90th percentile.


Yes, that would put a chill on DIY open source development. Would Dexcom or Abbott do this?

Ok still very good though :blush:

1 Like

quote=“Terry4, post:12, topic:64727”]
How often do you make manual basal rate changes running open loop? One or two per hour when you’re awake? That’s impressive that even open loop could reduce hypos so well.

With AndroidAPS you are required to complete a series of objectives as you progress towards the full closed loop with all the bells and whistles. Objective 2 requires you to run Open Loop over a week (you are allowed to turn it on and off) and make changes to your TBR as recommended. Under Objective 2 max_IOB is restricted to 0 so you are very limited in applying increased TBRs. According to the App, at the end of the 7 days, I had carried out about 130 changes in basal. However, the frequency varies a lot so after a meal bolus, it can be recommending Basal rate changes every 10 mins whereas overnight if your basals are correct, it might not do anything for several hours provided you keep your target range fairly wide.

AndroidAPS is really good at all of this,. There is a tab called CarePortal that allows you to log _everything_from carb corrections go site and sensor changes, cartridge changes etc. This links directly to Nightscout so everything is logged in. I did a sensor change and cartridge change this morning and both appeared in NS within a couple of minutes (and in the App). AAPS also links locally to Xdrip+ so the two Apps share data and calibrations entered in AAPS appear immediately in Xdrip+. The Careportal stuff has to go through Nightscout so it sometimes takes a while to appear. Because AAPS acts as a NS client, it will download stuff like basal profiles, carb ratios etc. from NS and then act on them locally on the phone


I assume you are aiming to build an OpenAPS system using the DanaR pump. Does that mean your three main components will be the DanaR pump, your Android phone, and a Dexcom CGM? In other words, you won’t need a radio bridge, like the RileyLink, to communicate with your DanaR pump. Is this right?