Accu-Chek Expert ...Active Insulin calculation

[JontyW]

How does the Accu-Chek Expert Meter calculate the remaining 'Active Insulin (insulin on board)?

This is what is shown in the 'Accu-Chek_Aviva_Expert_Advanced_Owner_Booklet', and I presume it uses the 'Acting Time' and 'Offset Time', but I cannot work out what the formula is they use, and Customer Services have not yet given me a proper reply (I was not impressed since they didn't actually answer my question).

Any ideas?

Acting Time
• The period of time from the start of the meal rise or the delivery of a correction bolus until your blood
glucose level is expected to return to the target level.
• You can adjust the length of the acting time to fit your individual needs, within a specified time interval
(1½§hours to 8 hours).

Offset Time
• Offset time takes into account the expected delay for the blood glucose level to actually fall during the
acting time of insulin in the body. It describes the first time period within the acting time.

Active Insulin
• Bolus insulin that has been given to lower your blood glucose, but has not yet been fully used.
• The meter automatically calculates the active insulin amount and it is displayed on the bolus advice screen.

 

[Sally71]

I think they keep the formula a secret! Supposedly any insulin bolused against carbs is not counted, only the correction part is tracked. But I found an instance once when it didn't work out correctly, I asked a Roche rep about it and she couldn't answer, just told me to trust it. Well yes, but I have a curious mind and would like to know what it's doing, if ever the meter goes wrong and I have to resort to manual calculations for a day or two it would be nice to be able to copy it!

Eventually I found a situation where it definitely didn't do what I expected, I emailed all the calculations to the rep and never got a reply

If I can find it, I'll post it here

 

[JontyW]

Thanks for the feedback. I don't have one of these meters since I do everything on a customised spreadsheet incl bolus calculate, but a member on another support site was also getting confused on what readings were coming up since there was no apparent logic to the Active units displayed.

The Active insulin remaining should be based on the '% Used' data that is available on the Novorapid/Humalog being used as a correction bolus. This is what I use but have no way of knowing how this matches my actual usage rate.

Hrs "Hum % Used" "Novorapid % Used"
0.50 4% 10%
1.00 20% 30%
1.50 45% 50%
2.00 64% 65%
2.50 77% 80%
3.00 85% 90%
3.50 91% 95%
4.00 95% 100%
4.50 97% 100%
5.00 100% 100%

 

[Sally71]

One thing I have noticed is that one hour after a bolus the pump will not allow another correction at all, after 2 hours you might get half (based on the 4 hour time for the Novorapid to stop working). I haven't changed the default settings in mine. And I understand that if your BG hasn't gone up more than the Meal Rise amount then you won't get another correction at all. But it's all very complicated!

Got to dash now but I'll try to think more about it later

 

[Sally71]

OK I found the email I sent to the rep. I'm not sure it helps, as I didn't note down the insulin on board, only how much correction the pump gave. I think what we can learn from this is that it isn't simple!!!

We have the Combo pump, the handset with it is the same as the Expert but with some extra programming to enable it to talk to the pump. So all the calculations should be the same.

At the time my daughter's ratios were as follows:

Carbs 0000-1630 1:18, 1630-0000 1:19
Insulin sensitivity (correction factor) 1:4 always
Target range in meter 3.7-7.0, therefore corrections are aiming at 5.3

On the day in question we had the following:

Breakfast at 1020, BG 8.7, 47g carbs, pump gave 3.4 units

47/18 = 2.61 units for carbs, 8.7-5.3 = 3.4, 3.4/4 = 0.85 units correction
Total 2.61+0.85 = 3.46 I agree with pump (I guess it rounds down)

Lunch time 1300, BG 11.2, 57g carbs, pump gave 4.0 units

57/18 = 3.16 11.2-5.3=5.9, 5.9/4=1.48. Total 3.16+1.48=4.64
Less than 3 hours since breakfast, pump seems to have knocked off 0.6 units (rounded down again) to allow for active insulin, how it arrives at this number I have no idea

At 1730 we did infusion set change, have to bolus 1 unit to prime cannula. I do this in bolus advice and enter 19g carbs so according to Roche rep this will not be counted as active insulin.

At 1740 tea time, BG 12.0. 76g carbs, pump gave 5.0 units

76/19=4 pump gave this + 1 unit correction, but at sensitivity 1:4 this will only bring BG down to 8
I make it 12-5.3=6.7, 6.7/4=1.675
Pump seems to have dropped approx 0.6 off the bolus again, but there should be no active insulin on board. Cannula prime was entered as carbs so shouldn't count, and the previous bolus before that was well more than 4 hours previously so should no longer be active either.

But I also know, from experimenting at different times, that if I had done one single bolus of 95g carbs (cannula prime + carbs together) then it would have calculated the correction amount properly and given a bolus of 6.6 units



I never got an answer from the rep about what is going on here!

 

[everydayupsanddowns]

Yup. Much to the confusion of pretty much everyone the Accu-chek Expert *only* tracks the CORRECTION part of any dose when calculating active insulin.

I had to call the lovely peeps at the Roche helpline when I was using the Expert to make sure I had this right as I was getting in a complete muddle with active insulin only reading 0.2u around 30 minutes after a full meal bolus.

I'm not sure but I think they use the ole reverse 's' shape curve in terms of estimating the insulin decay.

Any insulin that is calculated to be dealing with carbs is not included. The Expert doesn't have any of those prime settings to worry about.

If you want to track the *whole* amount you need to back out of the calculator wizard before confirming and then enter the insulin (to the amount the wizard suggested) manually.

Hope that helps
Mike

 

[Sally71]

That doesn't explain my example above - the pump doesn't know that I'm priming the cannula, I enter it as carbs so as far as the pump is concerned daughter just ate 19 g carbs so it shouldn't count that bolus as active insulin. Yet clearly it does because when I do another bolus 10 mins later with blood test it doesn't give the full amount of correction. But if I enter the whole lot together then it does give the full amount of correction! (which is what I tend to do now)

I can only conclude that there is something a bit odd about doing a bolus for carbs only without testing first!

 

[JontyW]

Sally,

Thank you for your detailed feedback with all the excellent example data which shows that you certainly seem to be well educated on how carb count & bolus adjust should work ... well done!

There are a few worrying things in your data, such as ...

-why does the pump round 'down' to show 3.46 as 3.4u, when mathematically it should round 'up' to 3.5? It is only a detail but it is not correct imo.
- "Less than 3 hours since breakfast, pump seems to have knocked off 0.6 units (rounded down again) to allow for active insulin, how it arrives at this number I have no idea". That is indeed worrying and needs Roche to explain what caused this.
- your daughter's BG readings prior to both lunch & dinner are on the high side at 11.2 and 12.0 against a pre-meal target of 5.3, so something is going wrong, such as pump not giving enough insulin. This assumes your carb count is correct.
- was this day 'typical' or a one off and normally she is in the target range?

Regarding Active insulin, I conclude that Roche need to make it clearer in their Advanced Owner Booklet exactly how the Active insulin is calculated and add a graphical example as they've done with all the other calculations.

Jonty

 

[JontyW]

Yup. Much to the confusion of pretty much everyone the Accu-chek Expert *only* tracks the CORRECTION part of any dose when calculating active insulin.

That is what I would expect it to do, only base it on 'correction' bolus. Since why would it need to take account of unused meal carb bolus which is doing a completely separate job?

Thanks for your feedback ... Jonty

 

[Flutterby]

This thread is very interesting, I too am questioning how the meter calculates the active insulin. It surely should be explained to us when we are given the meter that it only bases this information on any correction part of the dose. I've been very confused recently.

 

[Sally71]

Hi Jonty

Glad you found my example useful.

I never really thought much about the rounding up/down thing, a difference of less than 0.1 unit is negligible I think. Maybe it rounds down for safety reasons, I.e.so as not to send you hypo? But like I said, such a small difference surely wouldn't matter. Don't know. Must be some logic somewhere!

This data is all from summer 2013, I checked quite frequently and when doing a bolus at roughly the 3 hour mark since the previous one, it almost always reduced the bolus by 0.6. I would have thought it would vary a bit depending on the size of the boluses! Then again, my daughter is a creature of habit and does seem to eat the same meals frequently. I must admit I tend to just let the pump get on with it now and haven't checked its calculations for ages.

I didn't mention it in my previous post but you noticed that that particular day was not a brilliant one in terms of blood sugars! Great for checking what the pump was doing, but not very good for her. I'm glad to say that we do usually have better control than that, I think I had to do some basal adjustments just afterwards. Have been having the same problem this week actually, today we were mostly in the 8s rather than 12s which is an improvement, hopefully one more tweak will do it. As well as everything else, children have an annoying habit of growing, which usually mucks things right up just when you've got it working perfectly

If anyone out there does find out how the active insulin is calculated, I would be interested to know too. I suspect it's a closely guarded trade secret though!

 

[Annette]

Hi Jonty

Glad you found my example useful.

I never really thought much about the rounding up/down thing, a difference of less than 0.1 unit is negligible I think. Maybe it rounds down for safety reasons, I.e.so as not to send you hypo? But like I said, such a small difference surely wouldn't matter. Don't know. Must be some logic somewhere!

This data is all from summer 2013, I checked quite frequently and when doing a bolus at roughly the 3 hour mark since the previous one, it almost always reduced the bolus by 0.6. I would have thought it would vary a bit depending on the size of the boluses! Then again, my daughter is a creature of habit and does seem to eat the same meals frequently. I must admit I tend to just let the pump get on with it now and haven't checked its calculations for ages.

I didn't mention it in my previous post but you noticed that that particular day was not a brilliant one in terms of blood sugars! Great for checking what the pump was doing, but not very good for her. I'm glad to say that we do usually have better control than that, I think I had to do some basal adjustments just afterwards. Have been having the same problem this week actually, today we were mostly in the 8s rather than 12s which is an improvement, hopefully one more tweak will do it. As well as everything else, children have an annoying habit of growing, which usually mucks things right up just when you've got it working perfectly

If anyone out there does find out how the active insulin is calculated, I would be interested to know too. I suspect it's a closely guarded trade secret though!

Re:rounding. It is rounding each figure separately before the final addition-2.61 is rounded to 2.6, 0.85 to 0.8. This then gives 3.4 'correctly'. (Very poor maths skills there.)

 

[Sally71]

Re:rounding. It is rounding each figure separately before the final addition-2.61 is rounded to 2.6, 0.85 to 0.8. This then gives 3.4 'correctly'. (Very poor maths skills there.)

I wondered that too, but most people would round 0.85 up to 0.9, or so I was taught at school! That was a long time ago though...

 

[everydayupsanddowns]

Apologies Sally - I hadn't quite picked my way through your post when I replied. And you are right - there's something very odd going on there.

Wen I was using the Expert I pretty much gave up on 'Active Insulin' because I didn't find it at all helpful. It is frustrating that Roche can't give a clear explanation of how those results were obtained. It's a long time since I used the Expert so I can't really recall the effect of Meal Rise on those things - but if anything my recollection would be that the timeframes you are talking about rule that out too.

A mystery.

Jonty - There are clearly different viewpoints on Active Insulin - glad that it is what you expect. If Roche had called it 'Active Correction' I guess I would not have been worried, but when I check an hour or two after eating just to see if my dose/calculation/guesswork are proceeding as planned I find it more helpful for the whole amount of the dose to be tracked (as it is on the Veo) so that I can estimate whether the dose and correction are more or less on track. Testing less than an hour after a large meal with relatively high BG and only seeing 0.1u of insulin as being active never really helped me very much.

Just shows how different we are 🙂

 

[Sally71]

No probs EDUAD 🙂

I think it's silly that Roche either can't or won't explain this, I don't see why it should be a secret! I get the impression that the rep now thinks of me as a bit of a smart a*** who is asking too many questions, but I'm only curious, I'm not trying to tell them they are wrong or anything. And I only noticed this by accident one day when the pump gave exactly 1 unit correction but BG was higher than 4 units above target so it didn't seem to make sense and I decided to monitor it for a few days. Sadly I don't have the data for any other examples any more, I only found this one because I never clear my emails out 😱

The problem with bolusing without testing might well be a glitch in the software, but in that case I think they should be thanking me for pointing it out!

Or maybe it's something to do with the fact that it doesn't like you doing a bolus without testing and always reminds you that you should test first. Some silly safety feature perhaps? I don't see why that should be when I've clearly told it that that bolus is for carbs.

Oh well I don't suppose we'll ever know!

 

[Annette]

I think it may be to do with the 'false meal'you have to put in when you prime the cannula. If your meal amount (the minimum you can eat before it expects your meal rise to occur) is less than your false meal then it will expect/allow for the BG to go up by your meal rise. If you haven't tested, it doesn't have a start point to go on, so instead of giving the correction to get you back to the mid point, it assumes you are still within the rise caused by your meal (which you didn't have, but you told it you did), and will correct to the higher limit instead.
Im not sure I've been very clear there-I understand in my head but can't quite explain it!

 

[Sally71]

I think it may be to do with the 'false meal'you have to put in when you prime the cannula. If your meal amount (the minimum you can eat before it expects your meal rise to occur) is less than your false meal then it will expect/allow for the BG to go up by your meal rise. If you haven't tested, it doesn't have a start point to go on, so instead of giving the correction to get you back to the mid point, it assumes you are still within the rise caused by your meal (which you didn't have, but you told it you did), and will correct to the higher limit instead.
Im not sure I've been very clear there-I understand in my head but can't quite explain it!

Aha, Annette I think you might be on to something there! Yes that does (sort of) make sense to me, I have a vague understanding of what these things mean but not really about how they affect calculations. That might also explain why the 0.6 reduction seems to be so prominent in our calculations. Thank you!!

Daughter is at school at the moment so I don't have the handset with me, but I might have a fiddle tonight and check all those settings and see if I can understand them better!

 

[JontyW]

Hello All,

I've now had a response from both Roche Customer Services and also my local DSN, since in this area (NW Surrey & NE Hampshire) all new T1s are issued with Accu-Chek Aviva Expert meters when completing a DAFNE type course.

Roche Customer Services ...

"The active insulin displayed equals the full correction bolus during the offset time (expected delay for blood glucose level to actually fall during the acting time of insulin in your body). After the offset time has ended the active insulin decreases from 100% to 0% linearly until the acting time (acting time of insulin and the increase in blood glucose after the meal) has run out. Offset time and acting time can be set up individually for each patient."

My DSN says ....

"There is a pre-set algorithm in the meter. I suggest you speak to the Roche careline for the finer details of the calculations but it will make a deduction for any corrective insulin still in the system and this is based on the acting time. The factory setting for the acting time in the meter is 4 hours. When the blood glucose is elevated the aim is to bring down the blood glucose to the mid-range of the target."

So I conclude that after the Offset Time (1 hr), it uses a linear reduction in Active insulin over the next 3 hrs until the the Acting Time (4 hrs) is reached. So that is a period of 3 hrs (4 hrs - 1 hrs) when the remaining 70% (default Novorapid % Used is 30% after 1 hr) is divided equally over the following 3 hrs. But I don't understand why my DSN mentions anything about .. "When the blood glucose is elevated the aim is to bring down the blood glucose to the mid-range of the target"?

Any comments?

 

[Sally71]

That's what I was doing with my calculations - the Roche pump will aim to bring your BG down (or up) to the middle of the target range when calculating corrections. The default target range is 4.0 - 8.0 so it will aim to bring you back to 6. Therefore if you are 14 and have 1:4 correction factor it will calculate 14-6= 8, 8/4 = 2 units required to bring you back down to BG 6.
Our target range has been changed to 3.7-7.0 so it's aiming at 5.3 which makes tha calculation a little more complicated but it works it out in the same way.

I believe some makes of pump only aim to get you anywhere back within the target range so will only calculate to get you back to the highest point, e.g. 14-8=6, 6/4=1.5 correction units given.

Of course this only works if you are more than 4 hours (or whatever the acting time is) since the previous bolus. If you are still within the acting time then all that stuff about active insulin and meal rise comes into play and the calculations get a lot more complicated!

 

[JontyW]

The default target range is 4.0 - 8.0 so it will aim to bring you back to 6. Therefore if you are 14 and have 1:4 correction factor it will calculate 14-6= 8, 8/4 = 2 units required to bring you back down to BG 6.
Our target range has been changed to 3.7-7.0 so it's aiming at 5.3 which makes tha calculation a little more complicated but it works it out in the same way.

What is the point of having a target 'range' when all that is needed is a target 'value', which would be the mid point of the normal range? In my spreadsheet bolus calculator I just put in a BG target of 5.0 for daytime and 7.0 for pre-bedtime, and so all bolus calculations are based on getting BG to meet those targets.

So what is the point of specifying a target range, since it seems to over complicate it?