My review of Omnipod 5

I started with Omnipod 5 about two weeks ago. It was a bit of a rough start for my first few pods but now that I’ve got things figured out I am loving it. As soon as I switched from Novolog to Fiasp I got better results. Here is my graph of the past 24 hours, range 70-150:


Only a few Android phones and no iPhones can be used with the O5. That was not a problem for me…the Omnipod PDM is fine. It is an upgrade from the Dash PDM, it has a larger screen and a better screen layout which is easier to use.

The O5 definitely has some quirks that took a while to get used to. The main defining quirk is the way that the “Auto Mode” works. It dribbles out small amounts of insulin, every five minutes, according to its prediction of your BG in 1 hour. It does this by reading your CGM and trend directly, in my case a Dexcom G6. This is explained in detail in some publications such as this one: Review of the Omnipod® 5 Automated Glucose Control System Powered by Horizon™ for the treatment of Type 1 diabetes - PMC (nih.gov)

The greatest benefit of the Omnipod 5 for me is during those times when I don’t want to or can’t pay attention to my CGM. Sleeping and those gaps of 3, 4 + hours between meals. The Auto mode works great for making its own decisions about basal needs during those times as long as you bolus correctly at meals and correct manually when your BG trend is starting to go off track. If you go off track, Automode is not good at bringing down a lingering high BG and has no way of correcting a low. When correcting a high in Automode, it seems like correction bolus adds to IOB which then causes the algorithm to reduce the size of the auto basal causing a diminishing return over the next hour or so.

It sometimes counts part of the auto mode dribbles as basal and part as correction bolus, which seems to be based on where your BG is trending at each five minute interval. But it does this “behind the curtain”, resulting in IOB numbers that are hard to believe. In other words, you can look at the bolus calculator and it will report an IOB number that makes you say “How could that be?”. That IOB number was derived from a fraction of your Auto mode insulin over the past few hours. I often ignored that calculated IOB from Auto mode.

In Auto mode, it will go for long stretches (hours) of giving you zero or close to zero insulin. That happens if you are flatlining below the “goal BG” of the Omnipod 5, which can be set no lower than 110. For me, I had those long stretches of no basal in the early pre dawn times and often in afternoons. They caused a problem of breakfast bolus not working as it should have, with huge long lasting BG spikes. The way I solved that problem was to switch to “manual mode” about 30-60 minutes before eating. By making the switch, now I was back to getting basal according to my basal program. Manual mode also lets you use extended bolus, which is disabled in auto mode. Then about 60-90 minutes after eating I switched back to auto mode, once I was convinced that the CGM trend was headed the right direction. After struggling with the IOB calculations the PDM was reporting to me, I ended up using manual mode for all meals and I bolus just the way I used to with the Dash. That means I have to ignore the IOB in the bolus calculator when appropriate.

10 Likes

@John58 That’s a nice looking graph, particularly considering the range you state (I presume its Dexcom from the style, don’t understand their reluctance to put numbers on it so the graph can be interpreted. If anyone has a solution, I’d love to hear it.).

You mention my primary concern with the O5 and others: the “black box” approach. It may just be the geek in me, but I like to know what’s being used how; otherwise, it might as well be snake oil for what ails you. I don’t mind a “secret sauce” or “secret ingredient” to an extent, but there are limited factors that can be used and the equations are pretty well known…unless Insulet has come up with something new…highly doubtful. If I’m going to trust my life to these guys, I want to know what they’re doing. The other thing I’ve noticed in reporting: people either rave about their control or they hate it because of highs after meals…very little in between. That said, if it’s working for you and gives good TIR with minimal lows and its better than the alternatives, go for it!! I may join you in trying it out once the iPhone software comes out!

5 Likes

I think the Omnipod Auto mode is a forced fit for handling a meal. There are too many variables at play with the IOB, bolus and carbs that the algorithm does not respond quickly enough to. That said, the slow response of the algorithm also makes it feasible to cruise along in Automode at 80-90 all afternoon with a few well timed mini corrections.

3 Likes

I’m currently on the DEV branch of Loop so I can use DASH with my iPhone but after thinking long and hard, I bit the bullet and put in an order for the OM5 which is currently in pharmacist verification and should be shipping soon. I know an iPhone version is in the works, but what made me pull the trigger were things like:

  • fully supported vs DIY
  • 110 is good enough for me
  • Fairly easy to use
  • MOST IMPORTANT: direct communication between the pod and the G6 with the PDM only needed for bolusing and changes.
  • Gluko so my endo can see everything and not just rely on G6 Dexcom readings.

Once I get it, I will write reviews as well.

From what I can tell, it really helps to have your #s dialed in when you start, becaue the pod starts off using those until it learns, which can take 2-5 pods. So if our ISF and such are off, then it will be a rocky start.

5 Likes

@bpollina Best of luck with the transition! I look forward to hearing your report(s). As I’ve said elsewhere on FUD, I considered doing similar, but given my current numbers with Loop_dev and the requirement to either use a PDM again or change to Android, I’m just not willing to go that route. I will reconsider when the iPhone app comes out and gets settled based on reports from folks like yourself.

4 Likes

Yes; take the mobile out of the loop! Worth an acronym, TTMOOTL. Maybe not.

I’m doing AndroidAPS (with Dash) because I don’t want to kick the sleeping dragon and try to get O5 (no endo, dipshit PCP, good insurance, but why ask?) When thinking about reviewing it the thing that sticks in my mind is that I’ve gone from carrying around two pieces of electronics (iPhone, DashPhone) to just one! Yeah, half the searching, no; it’s better than that, one tenth the searching because I can remember where I put one thing down at least half the time.

Zero things; bliss, I can do quite advanced arithmetic, zero times any finite thing equals zero. No searching.

Maybe I should ring up my insurance company, ask whether I’m covered, persuade my aforementioned PCP to change the prescription to O5 (honest gov, it’s the same motor as the old one, they’ve just discontinued the old one, honest!) I’ll give [strikeout]Nurse Ratched[/strikeout] AndroidAPS a few more days.

3 Likes

im not sure how far along you are in your objectives or if you unchecked them, but i can verify aps works a lot better on the closed loop…the open loop kept asking me to verify temp basal changes, where the closed loop is auto…also make sure your profile is set up with a correct target…i started at 120-120 and wondered why it kept asking for temp basal corrections…i was wondering if anyone has their wearos running…i cant open the app on my wearis watch, so im stuck with xdrip on my watch and aps on the phone

2 Likes

I just setup AndroidAPS on my daughter’s Moto G Power 2021 phone. Up until this week, I was planning on getting her the Omnipod 5 (she currently is on Dash). I’ve been reading mixed reviews of the Omnipod 5, so that was enough encouragement for me to go through the AndroidAPS process. We just finished objective 3 yesterday of answering the 28 multiple-choice trick questions (good intentions, terrible execution by the designer/programmer). Now we have to wait a week to satisfy objective 4. No idea how long objectives 5-10 will take before she can actually do closed looping.

I just want to wait longer for more feedback/reviews of the Omnipod 5 to come out before we make the switch.

2 Likes

just be very careful on objective #6 as the low glucose suspend will not give a basal…i thought i had a few bad pods, or too much yardwork… but it was low glucose suspend…not sure how objective 6 is supposed to help…maybe forcing you to do manual basals…i kept going over 300mg/dl and after the 2nd pod change i realized low glucose suspend=no basal

3 Likes

I’ve actually got to #10 as of this morning and I moved to oref1 and OpenAPS/SMB (at #9). Your daughter should consider carefully whether she wants to do this after she has run OpenAPS/AMA for a few days; I found that my results (time in range and time within my target range, 100-120) were excellent with AMA. SMB is a bit scary because it is much more difficult to see the boluses, AMA is just the same as the Omnipod bolus calculator.

Objective #4 does not require a week. When I started I had requested closed loop and was responding to the temp basal suggestions. I did not change my targets to reduce the number of them; no way am I going to move my targets that much and I’m used to giving frequent correction boluses (which is all the temp basals are.) So I think I’d done more than 20 when I started #4,

Objective #6 is the show-stopper; it really did take 5 days and the days are counted at midnight. I have to admit I found the silent disabling of options a really weird and misleading approach, so I was up at about 2AM to get out of the LGS mode to a proper closed loop.

You will find that Objective #8 is painful, although I’m pretty sure there are no checks because it can’t tell if you used autotune. In my case I’ve had autosense enabled from the start, but given the silent-disable I assume it only started working after #6. autotune is useful, but it is not built in to AndroidAPS; I checked the whole of the source code. Using webautotune requires that you have NightScout and that it is publicly readable; setting the web site to default to “denied” simply stops webautotune working because it apparently can’t handle the ?token on the end of the URL.

Indeed, I almost abandoned the whole thing at that point. It’s the standard multiple-guess and I have never been able to do those. There are 15 possible answers to each question and a delay of 1 hour on changing a wrong answer so it is possible to get through the thing in 15 hours. Most, but not all, of the questions have the traditional “one clearly bogus answer”. When I got to the point where I was about to give the whole thing up I just removed the 1 hour delay; it seemed pointless spending all that effort then not trying it when I could see how to get past the test.

I didn’t have that problem at all; sure LGS means no basal when it kicks in but it only kicks in if BG is below the low end of the target (100mg/dL with my settings). OpenAPS/AMA does exactly the same thing; the problem I had with LGS is that, because it won’t do a correction bolus (via a raised “basal”) it has to be done manually, but that’s easy - just hit the bolus calculator icon then hit OK!

My problem with LGS was precisely that; it only does “positive” corrections, only raises the BG. That wasn’t much of an advantage for me because I hardly ever run low with Dash; I think I under-bolus then do corrections.

I admit it is unexpected that the Open loop makes correction suggestions (i.e. the misnamed “temp basal” correction suggestions) and the Closed LGS mode doesn’t; that’s weird, but it’s the same as Dash. In both cases we have to notice the high (possibly as a result of an alarm, but I don’t use those) and manually click a couple of buttons to invoke the bolus calculator and deliver a correction.

I also agree that forcing people to use LGS for 5 days is probably unnecessary, maybe it would be better to disable oref1 and limit the oref0 maximum basal change for a while.

2 Likes

Can someone please explain the “objectives” they are mentioning in this thread. What is that all about?

Is that for Omnipod5 or is that for Loop?

I have built FreeAPS and have it on my iPhone and I didn’t have to do anything like that.

2 Likes

There’s nothing like these objectives for Om 5

2 Likes

So I guess it is Android only. It just seemed strange to me because there was no such requirement to do it when building for iPhone.

2 Likes

@Eric I was similarly puzzled and looked it up. See Objectives — AndroidAPS 3.0 documentation
I glanced thru them and it appears to be a means AAPS builds into its programming to force people to go thru certain steps/processes to ensure their settings are inline with the program’s intent so the build, execution, are safe and the person understands the impacts of any changes to them. It may be well intended, but possibly a step to far for many.

3 Likes

i apologize, the objectives are for Android aps, and it helps to unlock features like mini boluses, etc…i mostly started aps because im used to xdrip and android aps has a watch app, however, i can’t get the watch app to work…always something… i figured id wait until the 5 is more available, i know pediatrics is having a hard time acquiring now, and aps works fine for me

1 Like

@TomH and @RogerType1’sType1’s explanation is correct, and @TomH’s link is too; that link is necessary reading for the spoilers that allow you to reach the goals to proceed to the next level.

The documentation isn’t clear on what gets unlocked at each level, but when you start the game you have something which is equivalent to the Dash PDM; it has a bolus calculator but no more. The system as a whole does, however, function like the O5 in non-auto mode because it offers full integration of the G6 (via the Dexcom app, which you have to “build” yourself) and the Dash pod via bluetooth. That’s definitely one up from Dash+OmnipodPDM because data only has to be entered once.

Level 0 does automate correction boluses, so maybe that is a slight difference from O5; i.e. it will alert you to the need (in its opinion) for a bolus. IRC it may be possible to turn on low-glucose-suspend (i.e. using a zero temp basal to correct for low blood sugar) at that point. The main problem is that you can turn things on in the UI but they don’t actually get turned on in the underlying OpenAPS.

The UI is also way, way more complex than O5 and AndroidAPS is definitely bloatware. I guess I could say that OpenAPS is bloatware too, but the AndroidAPS devs did choose to leave stuff out like autotune, they could have removed one or other of oref0 or oref1 or made them build-time options. (For those who don’t know “oref0” is OpenAPS1.0 and “oref1” is OpenAPS2.0; two very different versions of the OpenAPS algorithms used by AndroidAPS). They could have vastly simplified the UI and adapted it to the actual OpenAPS version being used.

AndroidAPS is also very assertive in a way that suggest the devs didn’t think very hard about diabetes stigma. It has a nasty habit of telling me what I may or may not do; indeed when I try to bolus for food when my BG is over 180mg/dL it pops up a window which pretty much says, “You can’t eat that now.”

IMO the relative complexity really is the difference between O5 and AndroidAPS. Forget the need to “build” AndroidAPS; that just took me half a day (even including the forced but unnecessary build of NightScout, required to get to higher levels in the game). The real difference is that O5 is set-and-forget with a minimal UI and all the consequences of that minimalism whereas AndroidAPS is a maze of twisty little UI features all interconnected that allow me to do almost anything yet are a nightmare to navigate.

I regard the O5 as a moped that will get me from A to B just fine with a little bit of care and attention and AndroidAPS as a Jet Pack.

5 Likes

OMG, that’s horrible!

I am happy with the current setup I have with the older pods.

I can either use the pods with the traditional PDM (which has the BG meter which I am very happy about) or use my phone and the Riley Link.

The switch to the newer pods would remove the Riley Link, but would also remove the old PDM. I definitely don’t want to do that.

I also like the fact that I can bolus or eat whenever and whatever I want, without the device telling me what I can or can’t do.

Since the Android app is like a game with different levels getting unlocked, here is a game for the Android developers!

Work hard at it, and when you get an A1C like mine, then you can tell me what to do…




Sorry! That kind of stuff ticks me off. :man_shrugging: Apologies for thread hijacking and so forth! :hijacked:

5 Likes

It’s kind of interesting what all you can/can’t do with different pumps/apps…and building yourself thru the DIY process broadens that field significantly! The bad side is you have to spend a LOT of time and effort trying them to find the pro’s and con’s and what fits “you,” the documentation is too different for each and too dense in others to do the comparison easily…and someone else’s comparison may not/probably won’t focus on what you or I would. In some ways, it would be nice to have just a Ford, a Chevy, and GM version, just three, to choose from; but that would also limit the capabilities afforded.

And with that further topic excursion…we’re back to your regularly scheduled Omnipod 5 review!

4 Likes

I’ll summarize my Omnipod 5 review in a few words: the jury is still out. It is similar enough to Dash (bolus calculator is different) to be more or less a wash. The Auto mode only works well for me (a) when I am already in good shape BG wise and (b) late night midnight to dawn it is good at adjusting basal to flatline at its specified range. So, considering my co pay for Om5 is identical to Dash, it is worthwhile to make the switch.

But for those shooting for much lower BG numbers or overall tighter control, it is just one incremental step towards that goal.

4 Likes

Mine is due to arrive today and I’m super excited!

I was using DASH and DEV branch so I could use my iPhone 12 Pro Max and no riley link or equivalent. That was nice, but FOMO set in, and since the copays are the same (free!) I thought I’d give OM5 a try. If I hate it I can always switch back.

The main reason I switched was I love the idea that they talk direct to each other without needing the PDM or phone. The PDM is only needed for boluses. I aslo like that it’s fully FDA supported. I know 110 is the lowest it can go, but if I could get to and stay around 110 I’d be happy. It is also much more set and forget, and I’m ready for a tech break. Eventually they will have a full app for the iPhone, and then I won’t need the PDA very similar to loop so that works for me.

I’ve also started a thread over at Diabetes Daily chronicling my journey on OM5. I also did one for Loop. Not many techies over there though.

6 Likes