I don’t know the controls that Omnipod 5 gives you, but I can make comments from basic principles.
Both pictures are from the evening time, some time after dinner, and they both show BG in the 70s and 80s. So I would suspect that there is some excess insulin from the dinner bolus or subsequent corrections, and this is tending to drag the BG lower than the O5 algorithm would predict. This could happen because of too aggressive an insulin:carb ratio for the meal bolus, or because of a too-short setting for duration of insulin action. Either one of those will trick the algorithm into believing that the meal bolus has been fully used up when in reality there is still some remaining active insulin that is pulling the BG lower.
I also see that in the second picture they delivered insulin when the BG was 62. This surprising action indicates that the O5 has a significantly lower safety threshold than alternative closed loop systems. I don’t know whether you can set the suspend threshold in the O5 or whether that is a hard-wired setting in the algorithm.
I’m not sure what the technical term for this is so I’ll try the best I can. I’m sure I will butcher this explanation so forgive me in advance.
Omnipod 5 PDM has a 5 minute delay when recording the automated insulin dose. The recorded automated dose actually corresponds with the BG recorded 5 minutes earlier.
So in your first picture the .05 was given when you were 86 not 70. To see the amount you received when you were 70 you would need to look at was recorded for 8:54pm.
On your second pic you received no insulin once you hit 62. That .05 you see was given at 76.
I’m not knowledgeable on algorithms so I can’t be of much help on the why of it all.
If I see a meal or correction bolus might be too much I will turn on activity mode for a couple of hours or so. Or if I think the algorithm is giving me too much of a correction on its own. This doesn’t always prevent the low but it can reduce the amount of carbohydrate I have to consume if done early.
The algorithm completely ignores Duration of Insulin so changing that won’t help. Ditto with correction factors and basal rates. It’s only used for manual correction boluses that you decide to give yourself. The only thing you can change is Carb Ratios and the target Blood Sugar.
The 5 minute delay helps explain some of it but it still seems overly aggressive to be giving insulin when the blood sugar is below the target that it is set to of 110. These are examples from my 9 year olds omnipod 5 and he has been on it for a little over 2 weeks so we are still trying to figure things out a little bit. It seems like as soon as he has a meal and his sugar starts to rise a little as would be expected with a meal and it starts dosing him aggressively it is like it complete ignores the insulin on board from the bolus he was just given for the meal then inevitable he goes low and then he is low usually going into the next meal and so we can not prebolus as much so then he goes up from the meal and the cycle begins again. It just doesn’t make sense to me to be giving insulin when he is low in the examples above to me. I know they aren’t what is causing the low as insulin takes time to work but I think they are making it more difficult to get out of the low. I know it has to learn but my son seems to have hypoglycemic unawareness as he never complains of low symptoms so it scares me for it to be so aggressive. I’m scared he will be out of range one time and get dangerously low without us realizing it. We have increased his duration of insulin to 4 hours. We haven’t really used the activity mode as he often needs more insulin for activities so I didn’t really give it much noticed but maybe we could try using it after meals?
Based on what Josie said above, the one adjustment you could try is to weaken the carb ratio so that the meal bolus is not quite so big. It does sound like there’s too much insulin at meal time from your comments about a slight rise followed closely by a low at mealtime.
What is the makeup of the meals? If it’s mostly carbs with small amounts of protein or fat what you’re describing can happen pretty easily. Or if your son is still honeymooning a bit as well.
For example I eat oatmeal with fruit in the morning. Before OP5 I would superbolus(take basal for 2 hours and add it to the meal bolus) then temp basal 0 for 2 to 3 hours. Now on OP5 I achieve this by setting Activity Mode after breakfast for 4 hours. Sometimes at the 2 to 3.5 hour mark I will still have to eat like 10 carbs. So not perfect but usually works.
Your question is valid. I thought the answers given were informative. Putting this in perspective, consider that .05 micro-boluses are the minimum I’ve ever seen, they act gradually over the next few hours, & they ceased after 2. Perhaps the algorithm forecasts dips to this level will be followed by recovery in future hours that would call for at least this level dose to ward off a higher rebound. It’s only about 1/3 of what I would have gotten from 24hr long-acting Basaglar/Lantus prior to OP5.
First I just wanted to say thank you for the suggestions and information as we are still just trying to understand how things work. I hadn’t realized that changed the duration of action for the insulin wouldn’t help. We did try activity mode for a whole day and that did help with getting out of the lows more easily. As far as the I:C ratio goes we are using the same we had for the omnipod dash previous, but maybe we will have to lower it since the Omnipod 5 seems so insistent on giving him more insulin even when low.
I do not know if resetting with lower basal will help since with our initial settings we only put in that he was getting 3 units of basal a day and about 2.5 units of that was at night. During the day he was just getting alternating 0.05 and zero units of insulin per hour.
I did double check the time zone on the pdm as that is what we are using.
Does anyone know if there is a site with good information on which settings are adjustable during automated versus manual mode that I could read about. Or a website that explains how the algorithm works since maybe if I understood it better I could manipulate the settings more to get what we need?
For now we are just limiting the amount of insulin the omnipod 5 can give per hour and just giving sugar to get out of the repeated lows hoping it will learn but I do not like being on edge all day worried we will miss a low and have something happen.
The algorithm does not distinguish day from night even when you program the lower daytime basal in. Just for the first two days. It should see when you are trending low and cut down basal during the day. But it will not learn that someone has a lower basal need during certain hours of the day. Even if they consistently need much less basal during the same time every day. This is a big flaw in the OP5 in my opinion.
I might try setting the daytime target to 150 in the settings for a day or so and see if it helps. At least the basal will be stopped a bit until the rise is over 150. Plus maybe look at adding some fat into the trouble meals. If you were doing well with Dash you can switch to manual mode and run it like the Dash. The only settings you can change that will make a difference are carb ratio, target bg and activity mode. That’s it. I guess you could play around with pre bolus timing as well.
You might check the tslim pump or DIY loop with Dash. These options would allow for the basal changes you need.
The three Juicebox podcast episodes explained everything the best. Your CDE was given a 5 hour presentation (at least mine did). It might help to schedule a appointment to have someone look at all your Dexcom numbers and OP5 data and suggest changes.