Top critical review
12 people found this helpful
Inexpensive but well short of the manufacturer's stated accuracy
on 9 February 2016
When I got the Codefree meter and strips I noticed there were a number of negative reviews that reported wildly varying results from consecutive tests done within a few minutes. I wondered if that was the norm and if satisfied users were simply unaware of a problem that could actually render the meter useless! To find out I took four consecutive fasting readings each day for twenty days using two batches of strips. The results are shown below.
In the leaflet supplied with the strips, the manufacturer states that a test they carried out with patients showed that 81% of readings were within 5% of the true value. Only 63% of the readings I took over twenty days were within 5% of the true value*. We can see from the results how the readings fluctuate. Every reading in a set of four should be the same! In addition to the fluctuation, the meter wasn't correctly calibrated and gave readings that were on average too high.
To put the results into perspective, let's take an example of a woman who is prediabetic** and on a particular morning has an actual fasting blood glucose level of 6.5 mmol/L. Very simply and roughly, there is a 50/50 chance that the Codefree reading would be in the range of 6.2 to 6.9. To express it another way, and more importantly, there is a 50/50 chance that the meter would give a reading that is somewhere OUTSIDE of that range. It is quite possible, therefore, that the meter could indicate to this prediabetic woman that she is "normal" one morning and has Type 2 diabetes the next!
My initial thought was to abandon the Codefree and find a better make. However, that might not be so easy! The latest industry standard for home meters requires that 95% of all readings come within 15% of the true value. For example, if your glucose level is 8.0 mmol/L an hour after a meal, the meter could indicate 6.8 on the downside or 9.2 on the upside and still be considered clinically accurate! The accuracy is permitted to fall even further for the other 5% of the time. As might be expected from such a low standard my Codefree was within it. A different meter could be significantly worse and still be within it!
The reviews suggest that the accuracy of a glucose meter can vary considerably not only between makes but also between individual meters of the same make. This is a simple test that can be done to see how well a meter is performing. Take two consecutive readings. Divide the highest by the lowest. Subtract 1. Multiply by 100. Repeat the test on several occasions. This can be at any time of day and on different days. Average the results. Ideally you want a figure of under 8. Under 6 will be equivalent to the accuracy claimed by the makers of Codefree. Above 12 and it's likely to be below the minimum industry standard***
At a figure of 9.7 my Codefree wasn't accurate enough to properly assess the effect of individual meals or foods on blood glucose levels. It simply wasn't possible to tell if an unexpected reading was the food or the meter! I now only use it occasionally for postprandial readings. On the other hand, the meter has been useful to assess the effect of changes in diet. I have taken two consecutive daily fasting readings and averaged them over a week. By comparing that figure week by week it is then quite easy to tell if the diet is making a difference to blood glucose levels. Using the meter in this way to monitor a trend over a longer term overcomes the calibration error and to quite an extent the fluctuation error as well.
FASTING BLOOD GLUCOSE RESULTS USING CODEFREE METER AND STRIPS
4 readings in mmol/L taken within a 3 minute period
Day 01 6.1 6.5 5.7 5.8
Day 02 6.7 6.8 6.6 6.7
Day 03 6.1 6.5 5.5 5.7
Day 04 6.3 6.0 5.7 6.1
Day 05 6.5 6.1 5.8 5.8
Day 06 6.3 6.0 6.7 5.5
Day 07 6.0 6.0 5.6 6.2
Day 08 5.5 6.3 6.4 5.6
Day 09 6.4 6.4 6.7 6.4
Day 10 7.3 6.7 6.3 6.2
Day 11 5.8 5.5 6.0 6.1
Day 12 6.1 5.8 6.3 6.4
Day 13 6.3 6.5 5.6 6.0
Day 14 6.0 6.7 7.2 6.3
Day 15 6.1 6.1 6.0 5.4
Day 16 6.2 7.2 6.4 5.7
Day 17 6.3 5.6 7.0 5.8
Day 18 5.5 6.4 6.3 5.6
Day 19 6.6 6.0 5.7 6.1
Day 20 5.8 6.7 5.7 6.3
*In the absence of a lab reading every day, and to put the meter in a favourable light, I took "true value" to be the average of the four daily readings minus the calibration error. 50/80 readings (63%) were within 5% of the true value.
** 6.1 - 6.9 mmol/L fasting is prediabetic according to UK health authorities
***This test assesses meter fluctuation only and does not take into account any calibration error that might also be present.