Regarding my diabetic meter's control solution, I am confused.

The solution is specified as being 0.12% (by weight) glucose, which I think translates into 6.67 mmol/l. However, the margin of error allowed for the machine testing the control solution is given as 6.7 - 9.0 mmol/l , which is all above the value I just calculated. It's supposed to be an equal plus/minus tolerance.

Where am I going wrong here?

[PS The actual reading for the control solution is 8.1 mmol/l]

Just to update:

I now have the control solution. It is described as a 0.12% glucose solution, which does equate to 6.67 mmol/l. However, I am told to be happy with a test reading of between 6.7 and 9.0, and indeed, it reads 8.1.

So what is going on here? The machine systematically reads high by around 20%. The manufacturers know this. But they don't tell me to adjust my blood glucose readings accordingly.

So all this time I have been reading my blood glucose as 6.6, when in fact it was about 5.3!

When you walk,
through the storm,
hold your head up high,
and don't be afraid of the dark,
At the end of the ....

....storm is a golden sky,
And the sweet silver song of the lark...

Lonely in here, isn't it.

I always wondered what purgatory would be like...

Try being a moderator...

Sorry - late to this. It's control solution not blood. Meters have to correct for all sorts of blood effects - haematocrit and other endogenous interferents. These don't exist within control solution. All you need to worry about is whether the meter reads within the specified range. It looks as though it reads nearly smack in the middle so that's good. Currently, under ISO standards, meter accuracy is only required to be (broadly) +/-20%. This is going to change.

Blood is incredibly complex stuff and varies wildly between individuals. This makes measuring glucose with a £15 meter pretty challenging.

Right - so the reading for blood should be correct when the reading for a control solution is as stated - around 8.00 for a 6.7 solution. I'll make that assumption.

But then I still have a problem. When I went to the hospital to donate a blood specimen for measurement, I took my own readings immediately before and after the hospital visit. All within the space of forty minutes or so. The two readings I took measured 6.3 and 6.7, but the hospital reading was 5.3.

You might say that my FBG level is capable of rising and falling quite significantly, so no problem, but at that time I was never getting anything like 5.3 at home. It was a complete one-off. It is very difficult to believe that I would have measured 5.3 at home when the hospital reading was taken. Much easier to believe that there is a systematic bias of around 1 mmol/l on my machine.

So, the bottom line to all of this is simply that I would like to know how accurate my meter is. That's all. I want to be able to compensate for any systematic bias it might be introducing.

Let me reframe my question then:

Is it possible for a machine to give the predicted reading for a control solution, but still introduce systematic bias when testing my blood? Could there be something within the complexity of blood that throws it out?

In short - yes, absolutely it is possible. Haematocrit, for instance, is definitely an endogenous interferent that could affect the reading of a blood glucose meter. Ascorbic acid (Vit C), uric acid and a wide range of others can affect the results too - all of which are naturally occuring in your blood.

Forty minutes is a long time but you wouldn't expect such a change necessarily between those two. It's also unusual for the bias ("built-in" offset) of a meter system to be high.

We always advise against comparing meter readings between meter systems but only compare against a lab reference system (Yellow Springs Instrument or similar). Do you know what the hospital used to measure your blood glucose? It's very possible that they used just another meter in which case you can expect anything up to 30% difference (even 40% could be acceptable).

What you can say from the control solution test is that the meter/strip is functioning as designed.

Feel free to PM me if you prefer - happy to help either way

Thanks -

The time between the hospital test and each of my home tests was only about 20 minutes, and also, as I say, at that time I was almost never getting anything under 6.00 at home.

The hospital is a major one (Glan Clwyd) and they had to measure all sorts of factors in my blood (lipids, HbA1c, total cholesterol, creatinine, etc., so I don't imagine they would have used a meter.

Ironically they probably use one of "our" (different part of the company)blood testers. Interestingly, I don't know how accurate these are for glucose measurements relative to a proper reference system - I'll ask though. Were most of the other blood measurements well within "normal" ranges? (of course, you don't need to answer that). If they were, it's less likely that this is the cause of the effect you are seeing though doesn't discount it.

Agree - I don't think the time will explain the differences though your liver can release glucose into your blood stream in response to effort. Generally, though, you need to be slightly cautious of results separated by any more than a few minutes.

Are you still using the Boots system? If so, let me know the model and I'll see if I can find out if I have anything on its bias.

http://www.journalofdst.org/September2012/PDF/VOL-6-5-ORG2-FRECKMANN.pdf

Accu-Chek Compact plus GT

Looks like any bias is minimal, and more towards the negative side.

I don't know how to explain the hospital experience. Either the hospital is using unreliable equipment or my machine is positively biased. Can you find out what they use at Glan Clwyd?

Re. My other results - cholesterol and LDL on the high side.

Need to be slightly careful with that particular study (Roche paid for it) but, yes, I expect your system is slightly negatively biased.

I'm finding out what to expect from hospital analysers.

By all means call Roche and ask them for an answer

So I have an answer and, to cut a very complicated story short, there is possibly a degree of inaccuracy in the hospital test too. It's very possible both systems meet their accuracy specifications but they just don't agree on the number.

Any idea which system they use, or the degree of uncertainty?

Don't know what system they use but seemingly these systems could be as much as 10% out on glucose so it's possible that the difference on reading that you are seeing is shared between the two measurement systems - hospital lab and your meter

Thanks a lot for all your help.

I'm discussing this with someone else on a diabetes forum, and after much thought and testing he tells me that his Freestyle machine reads 2.8 mmol/l control solution AS 2.8.

His other two machines read control solution AND blood too high, by the same margin. In all three cases, therefore, they do not discriminate between control solutions and real blood.

So as far as I can work out, the anomaly rests firmly with my machine. The control solution IS 6.67mmol/l, but the manufacturers specify a reading range of between 6.7 and 9.0 and, indeed, it does read as 8.1.

So either my machine discriminates in a unique way (unlike any other machine) between control solution and blood serum, or it just reads too high all round.

Our latest system automatically identifies control solution and deals with it slightly differently (it's pretty sophisticated in how it measures the blood)

I've replied to your PM and hopefully I'll have more answers tomorrow.

Do make sure you follow good practices: washing and thoroughly drying hands before testing, storing your strips in their container with the lid shut etc Given the relatively tiny quantities of blood used in these test, even the smallest quantity of contaminants causes errors

OK - thanks!

BTW my machine works automatically with presealed drums of 17 test strips. They don't get contaminated.