Blood Glucose Meter accuracy

[JohnMK23]

Hi

I was recently Diagnosed with Diabetes and someone bought me a Blood Glucose Meter (Sinocare Safe Accu2). I've been monitoring my levels first thing in the morning and two hours after my last meal. This morning the test was very high so I did another test which was very much lower, I did another two tests which were in-between. Being new to this I assumed these meters would be quite accurate but with today's tests being so different I'm not so sure.

So how do you get an accurate test and what is the best Blood Glucose Meter to use?

 

[harbottle]

We should have a sticky on the forum for blood glucose meter accuracy!

They are good devices, but some error is allowed under ISO standards, and over 5.6 the error margin is wider.

• Within ± 0.83 mmol/L of laboratory results at concentrations of under 5.6 mmol/L
  (Within ± 15 mg/dl of laboratory results at concentrations of under 100 mg/dL)
• Within ± 20% of laboratory results at concentrations of 5.6 mmol/L (100 mg/dL) or more

I've used a few, and decided to stick with Contour ones as I find the results from multiple readings are generally very close, and most of time both contour devices I use give me the same reading. (Contour Next has expensive strips, Contour Plus strips are about 10 pound a pot)

The Contour next is supposedly very accurate and is well within the guidelines.

Some of the cheaper brands I used could give me results that varied by the amounts in the guidelines.

So personally, I stick to Contour - I also like the App and Bluetooth connectivity, and results are uploaded to a website and you can get reports from it.

These devices don't work in lab conditions - the blood sample can have impurities in it, be mixed up with fluid from the skin, and the temperature can make a difference (the meter I use advices you to wait 20 minutes when moving rooms, so I guess it has a thermistor and does some sort of temperature compensation in its algorithm).

 

[Leadinglights]

I think it depends on the difference, monitors are allowed a 15% margin of accuracy but if very different it could be because you had not washed your hands. Many only look for trends and disregard the number after the decimal point.

 

[RBZ5416]

Also worth noting that meters are only required to meet those specs 95% of the time. So 1 in 20 could be wildly out but the meter (test strip) still be within spec. My experience with a Caresens Dual seems to match that & I just retest immediately whenever I get an unexpectedly high or low result. It's usually a high with the Dual & retesting the same blood sample can easily be 2mmol lower.

If you Google, there are a few group tests but mostly American. I did once find a UK one that included my Dual, but despite numerous efforts I haven't been able to find it since!

I believe Ascensia are still offering their Contour Blue meter FOC if you want to try another. Give them a call on 0345 600 6030.

 

[Docb]

To be a bit pedantic, it can be argued that what you are looking for is reproducibility which is different to accuracy.

The way I look at it is there are two things on the go here.

First is the intrinsic accuracy of the meter. You can only get that by repeated testing of a test solution, that is a solution where you know what the answer should be because the test solution has an accurately measured amount of glucose in it. To meet the standard for meters they should be within 15% of the known value of the test solution. I am quite sure that they will all meet that and most be well within that limit.

Second is the reproducibility of the blood sample. What you are taking is a tiny drop from a whole body full of blood and it is a bit much to assume that every drop will have identical blood glucose levels. If you take the drops from the extremities (fingers or toes) then you might expect them to have similar blood glucose levels but it would not be wise to assume they were identical.

Anyway my technical background kicked in when I was given a meter and after a while I decided I would try and find out just how reproducible the meter was and reported the outcome in another thread. I did 10 tests in rapid succession, one from each finger and then my two thumbs. I did a few stats on the numbers and found that there was a range of around 3 units from the 10 tests. This led to the simple idea that the number behind the decimal point is meaningless and from that day I vowed only to report blood glucose readings to the nearest whole number. I also decided that as far as I was concerned I would only react to a blood glucose reading if it was more than two or three units different to what might be expected. Even then, I would not panic, just do a bit more testing to make sure the odd reading was not "just one of those things" or actually suggested that there had been a significant change. It was how I got a good picture of the effects of different foods on blood glucose and latterly characterise the effect of stopping taking gliclazide.

Finally, in response to challenge by a member, I took a couple of tests from my big toe. There I got results which by my standards were effectively the same as those from my fingers. Can't remember the numbers off hand but rounding the meter reading to the nearest whole number they were not 2 units apart.

Hope that helps you get a bit of a perspective on the numbers.

PS.. Usual caveat, I'm really talking to T2's. T1 is a whole different ball game and as far as I can see, most T1's are a dab hand at interpreting the numbers, decimal points and all!

 

[Lilsis]

I have the same meter and initially had a few problems with it. I now test before my evening meal and 2 hours after and reading seem to have settled and give more consistent readings. As to how accurate it is, I won't know for sure until my next blood test in May.

 

[harbottle]

I agree with @Docb - consistency is why I use the Contour devices. The cheap model I bought was just a bit too random.

 

[Docb]

I agree with @Docb - consistency is why I use the Contour devices. The cheap model I bought was just a bit too random.

Your point is fine @harbottle and I am sure that there will be variability between meters and off brand cheap meters are likely to be the worst.

What I was trying to do was to introduce the idea that the biggest inconsistency might well be in the blood sample because it is such a tiny drop from a large volume. It could well be that this will limit the consistency you will get from any meter.

 

[Eternal422]

in response to challenge by a member, I took a couple of tests from my big toe

I remember that thread! I wondered if blood glucose varied throughout the body, but your big toe test confirmed that’s not really the case. Subsequently I read that blood circulates around the body in about 45 seconds (how amazing is that!) so I suppose that means that BG levels would very quickly become the same no matter where you took a sample from.

 

[Docb]

It was you was it @Eternal422🙂 - I could not remember!

I think that it is right that blood glucose would not vary wildly around the peripheries because of the speed that it moves about but I am pretty sure that if (heaven forbid) I rose to the challenge of testing all ten toes in rapid succession I would get the same range as with the fingers.

 

[harbottle]

Your point is fine @harbottle and I am sure that there will be variability between meters and off brand cheap meters are likely to be the worst.

What I was trying to do was to introduce the idea that the biggest inconsistency might well be in the blood sample because it is such a tiny drop from a large volume. It could well be that this will limit the consistency you will get from any meter.

Oh yes, the sample is key - as well as temperature of the blood, and the environment, and so on. Depending on the hardware and the technology, these could make a lot of a difference.

It started me thinking of a problem I've currently got on the product I'm working on - calibration of a clock driven by crystal. The cheaper technique works OK, but isn't fully satisfactory. As the process gets more accurate the costs increase, both in terms of new hardware we'd need and more complex software to manage the data and arrive at a value that's as close to the target as possible... (One variable is temperature, so one possible solution is adding a thermistor and trying to write some software that reads it and uses that the help the calibration.)

 

[everydayupsanddowns]

We should have a sticky on the forum for blood glucose meter accuracy!

They are good devices, but some error is allowed under ISO standards, and over 5.6 the error margin is wider.

This is posted in the ‘useful links for newbies’ @harbottle 🙂

BG meter accuracy
It can be quite disconcerting for members new to self monitoring of blood glucose to get different results from BG readings taken close together, even when carefully following manufacturers guidance (washing hands etc). All meters for sale in the UK should comply with the following ISO standards 95% of the time, which allows a degree of variation (and 5% of results can read anything at all). If in any doubt, or if a reading doesn’t match how you are feeling, you should check again with a fresh strip.

Contour XT meters and Roche Accu-Chek seem to get pretty solid results for reliability against lab results - though they are at the expensive end of the market.

At the more affordable end, forum members have found the Caresens Gluco Navii and Spirit Tee 2 reliable.

In fact I noticed that a paramedic from an ambulance was carrying a Caresens meter last year.