Dexcom algorithm for calculating IOB and meal bolus?

@Terry4 - I read the link but am still unclear on what the difference is between the Duration of Insulin Action and the Insulin Action Time. I also read the link within that but still it did not clear it up for me.

Also, what was your DIA before the change to 5 hrs? I am wondering how much of a change this was for you? As well, when you changed your DIA did you also have to make a change for I:C to maintain balance?

@Tim35,

The article is certainly a bit deep in the weeds! What I think itā€™s saying is this: the way the Insulin Action Times (IAT) are determined in a lab is different from the way that Duration of Insulin Action (DIA) is experienced in real life. In clamp studies where they measure IAT, basal insulin is suspended. Most of us are not routinely suspending basal to bolus for a meal (although some do, see ā€œsuper bolusā€). Hence, the DIA, how long the bolus is working in your body, tends to be longer than the IAT.

Do you have a Dexcom? If so, this is what you could do to get your DIA. In a period where you know that you basal rates are properly set (if you donā€™t eat during that time frame your BG will stay relatively flat), bolus for a meal and note the bolus time. Watch the decline of the curve after your meal, and note the time when it flattens out. Measure the time from your bolus to when the curve flattens, and this will be a good representation of the DIA you should use. My DIA is 5h and this has worked well for me. For the record, I use the Omnipod bolus calculator about 50% of the time.

One caveat: donā€™t run this test when youā€™re eating a larger than normal meal. There is some observational evidence that larger boluses have longer DIAs, so try to run the test when youā€™re having a typical-sized meal.

2 Likes

@Tim35, it looks a little arbitrary. DId you read the 2nd link? It might not help much, but, follow Terryā€™s link and then click the 2nd, here:

Itā€™s a confusing topic that Iā€™ve struggled with understanding for a while. I went back a few weeks ago and reread the John Walsh link as well.

When I started Afrezza I had read that it reached its maximum concentration in the blood in 12 minutes or so. I jumped to the conclusion that the 12-minute insulin concentration peak was also the maximum effect on glucose, but thatā€™s not true. The insulin concentration curve precedes the glucose lowering curve by at least 15 minutes in me. And Iā€™ll bet the two curves have different shapes.

During my pumping career, now 30 years long, my DIA number has varied from three to five hours. My recent change to five hours was from four hours. No, I did not change my insulin to carb ratio but perhaps I should have. Since I only consume about 50 grams of carbs per day, my I:C ratio is likely not as consequential as someone who consume more carbs. My I:C ratio is fairly aggressive, 1:4 midnight to noon, 1:6 noon to midnight.

Mixed up with this DIA topic are the terms pharmacokinetics (PK) and pharmacodynamics (PD). Hereā€™s a source that discusses those two terms in an understandable way.

Yes - I followed the second link. Unfortunately (and sadly) that did not help much either.

I did put an order into Amazon 2016 version of the Walsh book, ā€œPumping Insulinā€, sixth edition. Sounds like this is a very commonly read book on the forum but I have not read as of yet. Also ordered the 2015 Ponder book, ā€œSugar Surfingā€ in the same Amazon order. Between the two I assume I will get a good and accurate understanding of both these terms: Duration of Insulin Action and the Insulin Action Time as well as how to test for proper values for your pump in conjunction with wearing a CGM.

We had increased the DIA from 3 hrs to 4 hrs but were having high BG so decreased back to 3.5 hrs and it seems good there. But based on discussions, I am wondering if it should actually have been higher and the problem was that we did not also modify the I:C at the same time. Perhaps changing only the DIA without the I:C is not reasonable. I will read the books before trying another change.

1 Like

I dunno. I might just continue running the bolus off intuition. Iā€™m undecided.

1 Like

I do it that way. If you have years of experience, it can work fine. You will always be smarter than your pump.

Do you have good results doing it that way?

1 Like

I think so. I still ought to, at least, be aware of how the bolus wizard worksā€¦just 'cause. But, I feel fine, having stabilized my basal.

2 Likes

So I was not able to really understand the difference is between the Duration of Insulin Action and the Insulin Action Time. But based on CGM data, it appears for us to be about 4-1/2 hours from the time of the Bolus to the time when the CGM graph stops showing the downward curve which is likely from the bolus and goes into the level from our basal. (Many things assumed!!!)

So, we have changed the setting in the pump to 4-1/2 hrs and will see how it goes.

We also made some changes to the lunch, afternoon snack and evening I:C ratios to cause a bit more insulin to be delivered during these bolus meal times. Really on the fence about that but it seemed reasonable and we will be watching with the CGM to see if this seems more effective under various circumstances.

mohe0001, The User Manual has good calculation examples in the Appendix. Basically, for Omnipods, the bolus calculation is: Meal carb calc + Correction calc +/- Meal IOB +/- Correction IOB = Net Bolus

(OK non-Podders, I see your eyes glazing overā€¦)

It always delivers the full calculated bolus (or the amount that you override the calc) for the carbs you plan to eat now (carbs/IC ratio - Meal IOB). If you do not enter a BG reading in the bolus calc, then IOB is not used in the net bolus calc. It treats your meal bolus and correction bolus calcs in two tracks and both calcs are done with every bolus calculation if there is also a BG reading entered. Each (Meal vs Correction) has itā€™s own IOB tracking. A net Correction bolus (uses correction IOB if not zero) is computed when your BG is over/under your Target BG setting. It then compares that +/- amount to the Meal IOB. If the correction is less than IOB, then the proposed correction bolus is washed against the net bolus calc so you donā€™t stack up. Likewise if your BG was lower than your Target BG, the Correction calc will reduce the computed bolus based on carbs you are eating accordingly. The Length of Insulin Action setting is used to compute the IOB. Itā€™s a simple bolus amt/LOA. So a correction bolus calc will be > zero much sooner with a shorter LOA.

All in all Iā€™ve found the the bolus calcs to be pretty good. Exceptions: 1) I always bolus about 15% more on the first bolus after applying a new pod. I believe it doesnā€™t prime at much as it needs to. 2) I override the bolus calc to up it when I know darn well Iā€™ve eaten more surgary carbs than usual or that my meal carb guess was off. I have a 6 hr LOA setting so I force the correction bolus to kick in sooner.

1 Like

Thank you, thatā€™s real helpful.