Which is best, most accurate Glucose Tester

[Stuart60]

I ask this after doing a test with 3 different testers all with the same droplet of blood, (Yes I managed to get a really good droplet!) and got 3 different readings of 9.0, 10.1 & 12.5.
It is confusing to get a difference of 3.5 overall and given my recent accident which is currently being deemed to be due to a hypo as my reading an hour before the accident which was taken with the Freestyle meter of 4.7, a diabetes specialist nurse giving me a stern telling off for driving with a reading below 5.0 which given the ratio of difference I would have been over that using the AccuCheck device though equally even less using the Caresens?
Obviously this serious RTA which nearly took my life has made me very curious about these matters as I really do not want to be putting me or my family at risk

 

[Leadinglights]

I ask this after doing a test with 3 different testers all with the same droplet of blood, (Yes I managed to get a really good droplet!) and got 3 different readings of 9.0, 10.1 & 12.5.
It is confusing to get a difference of 3.5 overall and given my recent accident which is currently being deemed to be due to a hypo as my reading an hour before the accident which was taken with the Freestyle meter of 4.7, a diabetes specialist nurse giving me a stern telling off for driving with a reading below 5.0 which given the ratio of difference I would have been over that using the AccuCheck device though equally even less using the Caresens?
Obviously this serious RTA which nearly took my life has made me very curious about these matters as I really do not want to be putting me or my family at risk

Have a look at this thread about variation in results and the spread that any one reading could be.

What's in a number?

We have a lot of members, particularly those new to testing, asking about blood test results and what they mean or what is a "good" number. As many of you are aware I like numbers, mostly because they are what they are and do not have any opinion. This morning, I woke up, got out of bed, did a...

forum.diabetes.org.uk

 

[Stuart60]

Have a look at this thread about variation in results and the spread that any one reading could be.

What's in a number?

We have a lot of members, particularly those new to testing, asking about blood test results and what they mean or what is a "good" number. As many of you are aware I like numbers, mostly because they are what they are and do not have any opinion. This morning, I woke up, got out of bed, did a...

forum.diabetes.org.uk

Interesting read and looking on various websites including Diabetes UK it suggests the accuracy of a meter is allowed to be +/- 15% but the difference between the highest and lowest reading here if my school boy maths serves me right is 28%
Would it be normal to ask a doctor to get a lab blood test and before this going off dip test to see which comes back most accurate?

 

[Leadinglights]

Interesting read and looking on various websites including Diabetes UK it suggests the accuracy of a meter is allowed to be +/- 15% but the difference between the highest and lowest reading here if my school boy maths serves me right is 28%
Would it be normal to ask a doctor to get a lab blood test and before this going off dip test to see which comes back most accurate?

@Docb may be able to explain better as he has done quite a lot of comparisons with testing using different fingers etc

 

[silentsquirrel]

Interesting read and looking on various websites including Diabetes UK it suggests the accuracy of a meter is allowed to be +/- 15% but the difference between the highest and lowest reading here if my school boy maths serves me right is 28%
Would it be normal to ask a doctor to get a lab blood test and before this going off dip test to see which comes back most accurate?

If the "true" reading was 10.8 the highest and lowest are not much different from +/- 15%.

 

[Stuart60]

If the "true" reading was 10.8 the highest and lowest are not much different from +/- 15%.

But if the true reading was the 12.5 and not the 10.1 then the reading of 9.0 is well outside the 15% range as it would be if the 9.0 was the true reading I guess given the recent serious RTA I am likely being over cautious here and no doubt due to my PTSD-C over thinking things but worry I will do till I get a fuller understanding I guess maybe one I should be discussing with my PTSD Counsellor, though how much she knows about Diabetes I am not sure

 

[Leadinglights]

Your test may not have been consistent since it must have taken a few minutes to swap between all the three monitors so the drop of blood may have started to change.
I suspect you may have seen a difference with repeated tests from the same drop with the same monitor let alone a different monitor.
Expecting absolute accuracy from an instrument costing a few pound is perhaps misplaced.

 

[Docb]

Interesting isn't it @Stuart60.

Lets take the +/- 15% thing. What this says is that if you have a test solution of known glucose content then a test machine meets the specification if the results from that solution are within +/- 15% range. So, if you have a carefully prepared test solution containing say, exactly 10 mmol/l glucose, then a meter will meet the spec if it gives a reading between 8.5 mmol/l and 11.5 mmol/l. You have to be very careful when translating that spec range to interpretation of actual test results simply because with blood, unlike the test solution, you do not know what the blood glucose level actually is.

A second problem is what I call sampling error. Your analysis of the test results assume that each tiny sample from the blood drop has exactly the same glucose content. I do not have any information about the validity of that assumption and I do not have the kit to evaluate it but without evidence to the contrary, I would not rule out that there is some sampling error.

Where does this leave you? First, never forget that the reading you get is not an absolute value but an estimate of your blood glucose. So, my suggestion is that any number on a meter should be rounded to the nearest whole number. Indeed, since I did the tests reported in the thread "What's in a number", I have only ever reported single tests results to the nearest whole number. You can then add at least +/- 1 to that number to give you a range on your estimate.

The net effect is that the results you got could easily be due to statistical error and there is a reasonable chance that you will have got a similar range had you tested with one meter rather than three different ones.

The blood glucose meter is an amazing bit of kit and it is remarkable that it comes up with a number that is anywhere near the result you would get from a lab test on a decent size sample. You do however have to work within its limitations.

 

[Dragonheart]

The net effect is that the results you got could easily be due to statistical error and there is a reasonable chance that you will have got a similar range had you tested with one meter rather than three different ones.

I can vouch for this. I’ve tested the same drop less than a minute apart, ie as fast as the meter allows, on the same meter (accuchek mobile) and got variations up to +/-1mmol in the 5-7mmol range

 

[RBZ5416]

Just to muddy the water further, the standard that meters have to meet only prescribes numbers for 95% of the time. So one in 20 readings could be absolutely anything & the meter would still be compliant. About the best you can do is repeat any test that produces an unexpected result.

 

[Bruce Stephens]

So one in 20 readings could be absolutely anything & the meter would still be compliant.

The FDA apparently requires that the results ought also to be within 20% of the lab value 99% of the time (within the usable range). So likely for those 1 in 20 readings the result is not vastly out.

 

[helli]

If you are concerned by the accuracy of your meter, why don't you get some test solution?
This should be available from the manufacturers.

 

[Docb]

To throw in a further complication - the blood glucose meter does not actually measure blood glucose! I don't know how it works in detail but my guess is that it measures a voltage which changes as some chemical, impregnated in the test strip, reacts with any glucose in the drop of blood. Somehow, some clever stuff inside the meter, converts this voltage into a blood glucose taking into account all sorts of factors like the amount of blood on the strip, the amount of chemical in the strip, temperature, and anything else that can affect the reading. They really are remarkable pieces of kit!

At the end of the day the fundamental question is whether you can work with the out put and make sensible decisions. The answer is yes, after all Ti's have been using them since they were first introduced to good effect. Can you wind yourself up by trying to make sense of individual readings which are wandering about in the statistical error of the measurements? Again the answer is yes and is the reason used by NICE to suggest that meters should not routinely be given to T2's.

My suggestion is not to react to individual readings but to look for consistency and trends. To come back to the experiment which led to this thread. If it were repeated on several drops of blood and the three meters always gave readings in the same order, then that would be much more interesting and imply that there was a systematic difference between the meters. There is a pretty good chance that would not happen.

 

[Stuart60]

Your test may not have been consistent since it must have taken a few minutes to swap between all the three monitors so the drop of blood may have started to change.
I suspect you may have seen a difference with repeated tests from the same drop with the same monitor let alone a different monitor.
Expecting absolute accuracy from an instrument costing a few pound is perhaps misplaced.

All 3 were primed and ready to read as soon as I lanced myself, if they are not that reliable and thus open to a difference of over 15% means it is not really an accurate way for me to judge if I am safe to drive again and this will difficult for me to come to terms with due to my PTSD!

 

[Stuart60]

I can vouch for this. I’ve tested the same drop less than a minute apart, ie as fast as the meter allows, on the same meter (accuchek mobile) and got variations up to +/-1mmol in the 5-7mmol range

All my readings were within a minute and all were more than +/- 1mmol

 

[helli]

Thinking about this in the shower, this morning, I wonder if there is a little bit about our body's ability to get used to a certain level. Someone used to running their levels in the teens is likely to feel far worse at 3.5 than someone who is used to running their levels in the 5s and 6s. So, for a driving safety perspective, should the safe level be higher for the person with higher average levels? I know this is not possible to enforce and it is much easier to keep a fixed level.
But, if we apply this to meters that usually report higher or lower, someone with a higher reading meter will be as "safe" at 4.7 on their meter as someone with a lower reading meter reporting the same even if the true "lab accurate" is closer to one than the other. Because that is what their body is used to.
It is possible that the variations are too small to make that much difference but the advice to stick with one meter is repeated again and again as we can be driven insane by the variations between them.
We just need to remember the difference between accuracy and precision.

 

[littlevoice359]

For me, Contour Next One is the meter I take with me when I’m out and about. I find it accurate and the smaller size means I can fit all my gear into a zipped bag that fits comfortably in a jacket pocket.

The downside is that the test strips are expensive. But, now that I have the Dexcom G7 I find that I only need to test very infrequently - mostly to double check before a correction to a possible hypo/hyper and to calibrate a sensor after starting a new one.

I began to have doubts about the accuracy of the Freestyle meter I was given in hospital, so I started testing each box of strips using control solution. That confirmed my suspicions. I may have data on my computer still from an analysis I did at the time if anyone is interested.

 

[mikeyB]

On a more pressing subject, who decided that this accident was due to a hypo? What medication are you taking?

The reason I ask, is that (a) it is crazy to drive with a BG of 4.7 if you are taking any medication designed to lower your blood sugar. It's (b) also illegal if you haven't notified the DVLA that you are on such medication, so you may have been driving without a valid licence, and therefore without insurance, which is 6 points on your licence and a driving ban for up to 12 months.

That said, if you aren't on any medication, then you wouldn't go down to a level considered hypo.

If you have had hypos before, then you will inevitably be banned from driving - and you won't get your licence back without the DVLA thinking you're safe to drive.

 

[Dragonheart]

The reason I ask, is that (a) it is crazy to drive with a BG of 4.7 if you are taking any medication designed to lower your blood sugar. It's (b) also illegal if you haven't notified the DVLA that you are on such medication, so you may have been driving without a valid licence, and therefore without insurance, which is 6 points on your licence and a driving ban for up to 12 months.

not quite accurate

The uk law actually says don’t drive below 4. Between 4 and 5 before you set off you must take a snack to avoid a hypo if you use insulin. Additional rules apply if you hypo whilst driving too. https://assets.publishing.service.g...e-to-insulin-treated-diabetes-and-driving.pdf

If you are on other hypoglycaemic drugs you are still required to ensure you are safe - ie by avoiding hypos the same as for insulin users using the same testing guidance, even if you may not need to notify dvla. https://assets.publishing.service.g...vers-with-diabetes-treated-by-non-insulin.pdf It’s somewhat harder to find clear cut rules easily though about that need for testing. I’ve got them in posts elsewhere but haven’t got a chance to dig them out right now. I’ll add them again later.

 

[Dragonheart]

Ok. The leaflet above https://assets.publishing.service.g...vers-with-diabetes-treated-by-non-insulin.pdf deals with when you need to notify DVLA about your diabetes and medication - not about testing. It does say

“In the interests of road safety, you must be sure that you can safely control a vehicle at all times.”​

which would include not driving whilst hypo - especially when taking medications that can cause hypos. The only way to be sure you don’t do this is to test.

Then there’s the NICE guidance for type 2 NG28. https://www.nice.org.uk/guidance/ng28 In there is https://assets.publishing.service.g...84397/assessing-fitness-to-drive-may-2022.pdf
which says

“Managed by tablets carrying hypoglycaemia risk- Including sulphonylureas and glinides (for example Repaglinide, Nateglinide)​

May drive and need not notify DVLA, provided:​

■ no more than 1 episode of severe hypoglycaemia while awake in the last 12 months and the most recent episode occurred more than 3 months ago​

■ should practise appropriate glucose monitoring at times relevant to driving​

■ under regular review​

It is appropriate to offer self monitoring of blood glucose at times relevant to driving to enable the detection of hypoglycaemia.​

If the above requirements and those set out in Appendix D (page 127) are met, DVLA need not be informed.​

DVLA must be notified if clinical information indicates the agency may need to undertake medical enquiries.”​

so your dr should be both telling you to test and providing you the means to do so.


This post has more links to other documents that support this advice https://forum.diabetes.org.uk/boards/threads/very-high-h1ac-levels-practical-help.106752/page-3