Talk me out of going back from Om5 to loop dash….or not

Having all kinds of problems with Omnipod 5 and seriously thinking of going back to Loop Dash dev branch. Let’s discuss here.

1 Like

Ok Here goes. Sorry if this is long:

I started on Loop Eros with Orangelink. When the dev branch came out I tried the Dash and was loving it. But then OM5 came out and promised to learn me and ignore my settings after a few pods and talk directly to each other.

Well, after nearly 2 months of OM5, I don’t like it for several reasons.

  1. On loop dash, a spike was 150-175 and hardly ever in the 200’s.
  2. I have high insulin requirements, so I carried a FIASP pen with me and used that for bolusing.
  3. As a result, I was able to stretch a pod to nearly 48 hours or longer. Under OM5, I’m forced to not use a pen at all and can barely get 24-36 hours of use.
  4. When I’m out and about like today when I was at the theater, I went to bolus 30 units that I needed, only to hear the <10 units left alarm. Since I had no pen, I was forced to stop the bolus and go high. This wouldn’t happen on my old hybrid system.
  5. Under loop, my last A1C was 5.8. I’m now creeping into the 6’s.

I asked my endo what he thought he said I’m getting older and he’s more concerned about lows than highs. He said the OM5 is much better at preventing them, landing you softly. He also said that it’s okay to go into the low to mid 200’s so long as you don’t stay there. I’m not sure I agree with that.

Can we have an honest discussion about the pros/cons so that maybe others in the same boat can benefit as well?

1 Like

I use lantus 10 units a day to take the pressure off my sites.
It looks me to fly under the 112 limit on my pump and it also keeps my sites happy for 3 days.
I have a problem after 50 units in one site per day. And I generally use 50 a day. Taking 10 lantus means 40 per day that is pumped and it makes a big difference.
At night I run about 95 on average and I don’t crash
It might be worth trying since you are using a pen anyway, maybe a long acting pen could help.


Thanks for sharing. This is what I expect from a FDA-approved device. Idiot-proof, but leaving smart people behind. I am determined to stay with DIY Loop as long as I can. I don’t have insulin resistance yet. My A1C is 4.9. Why would I give up living a healthy life only to please some idiots. It is enraging to me that I have to pay these idiots and that they have such control. The government’s role should be to protect me from others, but not from myself. I should have the right to set my BG target to a heathy, normal level. It is OK to buy an assault rifle but not to keep my BY in normal range! I am beyond myself when I think about it. That being said, the US is still the cleanest dirty shirt. I am not up to fighting the system. I just evade the system and live happily ever after.


I don’t. I’m using AndroidAPS (which is OpenAPS with an interface I have criticized elsewhere). This is with the Dash pod. The O5 is attractive because it removes the 'phone from the loop and I don’t use external boluses but also I don’t use extra boluses with AndroidAPS; I just trust to its algorithm

My results have depended on my carb intake; I’ve increased my carbs recently and have spent more time outside the 70-180mg/dL range. My current stats report between 0% and 5% high (over 180) and my 7 day average is 2% high; I’m reducing carb intake in the hope that it will pull me back to the 1% of my 30 day average.

As for lows, well, below 70 0% over the last 7 days. I check with a fingerstick regularly, maybe religiously, below 80 and the main problem is that my G6 CGM consistently reads low. Particularly at the start of a sensor it can easily be 20mg/dL below the reading that my body and a fingerstick tells me, more at lower BGs.

So with me and AndroidAPS there is absolutely no problem about lows.

As for over 200. Well, I’ve posted a few times on this list above that and I was a truly evil bastard. I apologise.

This is about your endo’s comments. My mileage with loop is similar and different to yours; my BG spikes are similar yet I’ve used 28IU/day over the last 7 days. Since I started with the Eros pods (not looping) I’ve never had any problems going 80 hours, I typically put 150IU (half a pen) into the pod and I only exhausted the reservoir once or twice.

So my insulin requirements are far lower than yours and, more important perhaps, far lower than the pod can deliver; your cons for the O5 wouldn’t matter for me, but you don’t document what your spikes are with the O5. You seem to imply that they are bigger than Dash loop but it’s not clear.

I re-read my post and realized that it is inappropriate to call the FDA enforcers idiots. They do their job. They are being penalized for any downside and not rewarded for any upside. Ideally they should not approve anything at all. I should applaud them for approving the OmniPod 5. It would have been much safer for them to not approve it at all.


Almost every doc will say the same thing. Except they believe every diabetic should place a premium on staying out of hypoglycemia whether you’re getting older, are young, pregnant are working out or a most every demographic or lifestyle!

I hesitate to spell out my philosophy cause I’m sure to offend someone. I believe we are all entitled to normal blood sugars, whether we are diabetic or not. It’s tougher for diabetics but not impossible.

We live at a time when several tools may be exploited, if you so desire. I use a low carb diet with plenty of animal protein. I use the DIY Loop system with an old MiniMed pump and a Dexcom CGM. The algorithm resides on my iPhone and I refer to blood sugar levels and trends often.

Some people don’t like this amount of vigilance but I don’t mind. I think it feels freeing; it’s like more work leads to less.

My time-in-range 65-120 is often at 90% with less than 0.5% less than 54. My A1c is usually in the high fours or low fives. I think TIR is more important than A1c.

I correct higher trends with Afrezza 1-3 times per day. My total daily dose of insulin is about 30 units per day and my weight is steady at 155 pounds.

Everyone needs to strike their own balance. I’ve found mine but I know others can find their own.

I think you already found your solution.


Thanks everyone for the constructive dialog. Let me give some examples:

OK so here is my Dexcom Clarity Data:

01/01/2022-03/31/2022 (Loop with Dash Dev)
Average Glucose 131
Standard deviation
GMI 6.4
TIR: <1% very high
6% high
93% in range
<1% low
<1% very low
target range 70-180

Average glucose 125
32 std dev
GMI 6.3
<1% very high
5% high
93% in range
1% low
<1% very low

6/8 (start of OM5)-8/16 (today)
Average glucose 139
standard deviation 33
GMI 6.6%
<1% very high
10% high
88% in range
1% low
<1% very low

So while it is not very much of a difference, there is a pattern upward of highs and a slightly elevated GMI plus std deviation is a bit higher.

The thing is that the OM5 has a LEARNING algorithm that learns based upon total daily dose, so if I supplement with a Fiasp pen like I did with dash, it won’t learn properly, or so I’ve been told.

It also runs out quicker. I’ve also noticed that it makes it harder to find new spots on my stomach to put the darn thing because it typically takes a good 48 hours for the spot it came out of to fully heal.

I really thought I’d like the fact that the OM5 and the Dexcom talking to each other would be a huge benefit. But I almost always have my phone with me, so really its not as big of a deal as I thought it would be.

I’m still torn.

For you, the upside of DIY Loop seems to be about a 10% reduction in average BG, quite impressive! Also can set desired BG level, can input external insulin sources, can use external BG sources via Apple Health, and can increase your pod life. The downsides are annual Apple Developers fee of $100, you have to rebuild the app once a year, you don’t have a 1-800 number to call to solve your problems, and Apple obsolescence after 6 years your phone or computer can’t be used to loop. I’m 67 and have been looping for 5 1/2 years. I spend very little time managing things. The odd hiccup is easily solved. You have a new iPhone and new M1 computer, so I would stick with Loop


I think that is something you should consider testing for yourself. (My opinion, not backed up by anything more than gut feel is that the “learning” is pretty overrated, I think it learns pretty simplistically). If you could get 2+ days per pod with BG and TIR that satisfies you, would you stick with Om5? That question can only be answered by giving it a try.

Well, slightly off topic, but that is a “pro” for the Loops over the O5; getting the data out of the O5, or Dash or Eros or the previous version, is pretty much impossible. The only data aggregator available is Glooko. To make it worse log in on Dexcom web sites is total garbage; Dexcom has multiple sites with different log ins.

On the other hand the Loops are really only set up to handle NightScout, at least I was not given any choice with AndroidAPS, I had to set up NightScout. Nevertheless the data remains ours; we don’t become unpaid contributors to corporate research programs.

On your actual data I can compare what I’m getting with Dash+AndroidAPS. To do so I went to Clarity/Settings tab and changed my “low” and “high” threshold values to 70…180 (four changes). My normal range setting 80…150. Notice that the “target range” on the setting page cannot be changed; it is just a quote of “ranges from an international consensus”. This is very confusing :frowning:

Person Algo Period GMI AvgBG TIR TARG(mg/dL)
BradP LoopDev Q1 6.4% 129 93%
BradP LoopDev April+May 6.3% 125 93%
BradP O6 69 days 6.6% 138 88%
jbowler Dash PDM May-22 7.2% 163 60% 100
jbowler AndroidAPS/Open 14 days 6.9% 150 77% 110
jbowler AndroidAPS/Closed 30 days 6.1% 117 98% 100

I used “GMI” to back-calculate the corresponding mean BG because it is not clear what the Dexcom page describes; it doesn’t define what average of what they are displaying! It does look like a time weighted average of BG though; my back calc gives numbers that are the same (to 2sf).

I’m comparing a fully developed T1 (me) with a highly insulin resistant person (insulin resistance is one variation that gets glommed into T2 along with other things.)

One thing that does strike me immediately about the @BradP numbers is that compared to my transition from the open loop Dash PDM to the open loop AndroidAPS (oref2) then to closed loop (SMB, oref2) AndroidAPS my improvements are massive! This has been noticeable to me from the start.

The second point is something I noticed with AndroidAPS. I started with the default TARG setting, which is actually 105-115mg/dL IRC. I found that my BGs were, as might be expected, running higher and I was having spikes over 200mg/dL so I dropped back to 95-105mg/dL My understanding is that 110mg/dL is the lowest TARG possible with the O5; going from 100 to 110mg/dL raises GMI from 5.7% to 5.9% ( a raise of 0.2392% to be exact.) The @BradP raise from Q1 2022 to the 69 days of 05 usage is precisely that. Is this all that is happening?

Still, a definite “pro” for the Loops; moving from Dash to O5 means accepting a restriction on the target BG.

if I supplement with a Fiasp pen like I did with dash, it won’t learn properly, or so I’ve been told.

It’s much worse than that. The Loops are insulin-delivery agnostic; they take CGM data first and then suggest or implement insulin dosing. As a result you can take insulin any way you like. The Insulet Omnipod systems, all of them, are fundamentally pump controllers, O5 adds integration with a CGM but that is all. It has never been possible to record insulin delivered from another source and I doubt Insulet will ever implement that (it adds so much complexity to the testing).

This comment applies to all the pod systems when used with the Insulet PDM; Eros and Dash can be used with Loop but it takes over from the PDM, so the Insulet algos, particularly the IOB calculation, are replaced by the Loop ones. If you use the PDM or the O5 and bypass the pump by injecting insulin then the IOB calculation will not take this into account and subsequent bolus calculations will be completely wrong for the duration of the insulin action time of the injected insulin.

So there is no choice; the pods can be used for insulin delivery alone by not using the PDM bolus calculator, but then the actual calculation has to be done elsewhere (e.g. in your head, like the good-old-bad-old days of MDI).

There’s another case where something like this happens. For people who are not fully developed T1Ds, particularly people in the stablization period post-honeymoon, and for many, maybe even most, T2s who still have some semi-functioning beta cells the IOB calculation of any closed loop system is going to have much higher errors.

This is not primarily because of endogenous insulin production from the beta cells, it is because of the continued regulation of the alpha cells that cohabit the islets in the pancreas; the alpha cells regulate the release of glucose from the liver and this is what causes the requirement for a basal. The closed loop algos simply don’t know how much this is happening and factoring this in is very challenging, maybe impossible.

It is conceivable that a true learning algorithm could would out how basal requirements change with blood glucose, but the mechanism of alpha cell regulation is still not understood:

(I posted more discussion of the above article in another TUD thread.)

I’m pretty near certain the O5 doesn’t even attempt to do this. There may be some work in Loopland but I don’t know of it. It’s not a pro or con of either system, just a limitation on the accuracy of all systems for everyone except fully developed T1Ds.


im using Android aps with lyumjev u200, although i had help setting it up with the diabetes nutritionist/educator, and would easily run out if i was still on u100. u200 does seem a !

little slower, but something to seriously consider. im mostly using aaps because of the watch, which also shows predicted values


Sometimes we all get too caught up in the numbers. You had excellent control with the Loop DASH at 93% TIR, but you are only down to 88% TIR for the O5, so both are still well within the “control” range. You were in the A1C 5 range but are creeping into the low 6s. That still is considered excellent control! So either way, you are doing a great job controlling your diabetes. The occasional high or low occurs because you are a HUMAN BEING, not a machine. We all have to live meaningful, active, productive lives, and sometimes that means losing a tiny bit of control.

That said, the questions – which only you can answer for you – are these:

  1. What system (Loop DASH or O5) best serves your diabetic needs?
  2. Is the direct communication from CGM to O5 that much better and worth the slight reduction in fantastic numbers worth it to you?
  3. With which system are you most comfortable?
  4. Which system is more cost effective for you?
  5. Which system wastes the least amount of product and materials?
  6. Which system is the least intrusive on your daily life?

We all are unique in our needs, in our definitions of “good control,” and in our lifestyles. I hope your answers to the above questions will help you determine what system will work best for you at this point in your life. I trust that whichever system you choose will be the right system for you.


Thanks everyone for all the constructive criticism. After crunching the numbers I’ve decided that at least for now I will stay with the omnipod five because it is indeed very easy. Plus right now I didn’t realize I was out of my developer subscription and $100 is hard to come by right now. There’s also more headaches if I run into problems. And my numbers really aren’t all that different. I will keep a close eye on everything, and if I start going into the high 6 to 7 a1c then I will become more concerned.


As an aside has anyone tried ? I use it to get my glucose numbers in real time on my Apple Watch as well as they have it in a separate browser at work. So I’m even use the call me feature that literally calls your cell phone when it reaches a predefined either low or high threshold. I haven’t turned that on but some love it.

1 Like

You do have another option… How stuck on Apple are you? You could probably get an Android phone for free, as lots of people have old phones laying around. Just ask in your local community (Facebook group?). AndroidAPS doesn’t require a developer license and is free to build. Plus it’s got some bonuses not available on various Apple variants. It’s most similar to OpenAPS (super micro boluses, unanticipated meals, dynamic ISF, etc…) I’m not familiar enough to know what the bonuses are on AndroidAPS, not available on the Apple variants, but I see it mentioned often enough that there are things.

I’m all set up to run AndroidAPS myself. Have the app built, and first 5 training objectives completed with a virtual pump, and ready to connect it to my Dash pods… Just haven’t actually done it yet! I’m half excited, and half nervous. I’m scared to give up my Control-IQ, since you have to run through a few more objectives in open loop.


I found this interesting in a review of the Omnipod 5


dont forget about the wearos watch build for aaps, its a different app build, hut virtually the same as your original app…makes a world of difference especially during work or driving… it is also a little tougher to install now, but worth it. i would also imagine you have your xdrip ringtones/notifications all customized and set up also, but it seems like these are major advantages over omni5 and tandemciq…although tandem might have a watch app now, i haven’t used it personally

1 Like

Ha! That’s funny! Tandem barely has a mobile app at the moment. Okay, that’s not entirely true, most phones can run the bare bones app that shows you pump data/info. They can’t do anything with it, though. It’s one way communication.

There are only a handful of Android phones they’ve added mobile bolus capability for. The current list only has older flagship Samsungs on it. And that’s it, only bolus and view pump details. You can’t interact with the pump in any other way, including dismissing alarms/alerts, which I do WAY more often than bolus. I know this is the very first software version and other improvements are coming, but I feel like this is a massive oversight, or lapse in judgement given just how often the pump likes to alert. Especially if you spend much time under 100 mg/dl.

I’m guessing they’re still a looonnngggggghhh way off from watch control. I doubt they’ll take that on until after Mobi and it’s full mobile pump control are launched.

Do the Apple variants not have watch apps? I thought Apple watches could generally access just about anything on the phone, so thought this would be a feature of looping with Apple.

But excellent point about Xdrip! I’ve long valued it as the best CGM app, way before I ever even considered DIY looping. The Dexcom app is just painful in comparison. They still don’t even have a widget! (At least not in whatever version the BYO app is based on.)

1 Like

aaps does a pretty good job microbolusing unannounced meals, but its a lot easier just to add a few units from your watch… wear installer is key, and i personally had to reset the modem, router, phone, watch in order to install it…sad that diy is the only way to go, but at least its available… the article was hilarious on the advantage of diy, and they were only talking about Apple’s version, which seems less automated…ive watched many videos recommending insulin sensitivity changes for meals… sounds awful

1 Like