Living with the Libre – the bad

By | 20 March, 2018

I wrote about our second favourite flash glucose monitoring device – Abbott’s Libre – and covered off all the undoubtedly good things about the device. But all is not milk and honey in the land of the Libre and so I’ve jotted down a few of my annoyances for your delectation.

Some of these are practical day-to-day things and some are more esoteric and aren’t really Abbott’s fault but I’m going to lump them all together because no-one ever said life was fair.


A picture of a Libre shamelessly stolen from Abbott's website. This is copyright infringement in action, kids

A picture of a Libre shamelessly stolen from Abbott’s website. This is copyright infringement in action, kids

There’s no escaping the fact that any flash glucose system is not going to be as accurate as a capillary blood test. Prior to starting with the Libre, I used Bayer’s (sorry,¬†Ascensia Diabetes Care’s) Contour Next Link meter, which I think is the best and most accurate meter I’ve ever used. It’s a great bit of kit and you just get used to reliability and accuracy. Check twice in quick succession with the Contour Next and you’ll get the same result.

Everyone goes into using the Libre with their eyes wide open – we all know that interstitial readings will lag behind capillary. We all know it could be a bit wonky. But, in practice, I want accuracy – I want to know what’s happening with my glucose levels now and I want the results to be accurate, dagnabbit. I’ve compared the Libre’s results to my Contour Next and it’s not uncommon for them to be way apart, even when nothing much is happening blood-glucose wise.

I know I’m asking for the moon on a stick but my Libre / meter is my only source of data for deciding how much insulin I put in. Getting it wrong makes my BG go a bit high or a bit low. Mild inaccuracy is not going to kill me (calloused fingers crossed…) but it’s annoying to have to ;

3 thoughts on “Living with the Libre – the bad

  1. Richard Blogger

    Accuracy. How do you know its not accurate? The only way to find out is to compare the Libre with a *lab* test. Comparing Libre with a blood stick test is a bit pointless because they only have to be accurate to 20%. I have recently been changed by my GP to Accu-Chek Mobile (their decision) from Contour Next. I find that the Contour agrees regularly with the LIbre to within 0.2mmol/l, the Mobile can differ from the other two by 0.8mmol/l. So which is inaccurate?

    Interstitial. What’s wrong with that? You use glucose in your muscles, it is useless (actually, dangerous) if it just stays in your blood. When you measure interstitial you are measuring the glucose that is available to your muscles rather than the glucose that will be available in a few minutes time. Guess which I prefer? Imagine going to the supermarket to buy a can of baked beans, which is more useful to you: being told that there are cans on the shelf (interstitial), or being told that there is a lorry turning up in ten minutes time with cans (capillary)? I have been doing blood tests for 30 years (yes, I started with the original BM sticks), so I have to change how I interpret the value, that’s all.

    Reliability. I have used 14 so far and (fingers crossed) so far none have failed. I had one replacement because I got caught in rain, soaked tothe skin and the adhesive of the sensor failed and the sensor came off when I took off my t-shirt. Now I cover the sensor with Tegaderm. Customer Support was good, they asked a few questions and sent a replacement, asking me to return the sensor. The interesting thing is that if you buy sensors from their website it is pretty random when you’ll get the parcel, usually it takes 3 or 4 weeks (it comes from France). If you request a replacement it turns up in a few days because it is dispatched from England.

    “WHY PRICK WHEN YOU CAN SCAN” I agree with yopur irritation over this. I want to reply “in 42 years of T1D I have injected myself 50,000 times, why are you making so much fuss about finger pricking?” Using Libre is different to using blood sticks. The great power of Libre (and other CGM) is the trends. I am not talking so much about the arrows (althiough they are useful) but the 8 (or 24) hours of historic data. If I see my glucose rising after a meal I know to scan 45 mins later for the peak (-> arrow) and hopefully a fall to something in target. (On a rare day when I get the pre-bolus right, there may be no change in glucose!) If I don’t get the peak, my glucose continues to rise, then I can take a correction. I have learned that for me the rapid rises aren’t the problem, the slow rises are. If I get a -> reading that actually means my glucose is changing at less than 0.06 mmol/l per minute, which could mean it is level. However, in my case, I find often in the afternoon my glucose rises slowly so -> could mean a slow rise of 0.05mmol/l per minute or a 1mmol/l rise in 20 minutes (put that way, its not that slow, is it?). So even with a -> I scan again in 30 minutes and if my glucose has risen I then need to decide if I need a correction to stop the rise or not.

    The real bad? The reader (and the desktop app) cannot handle putting the clocks forward. In fact why doesn’t the reader do this automatically? People who travel a lot find the reader’s problem with times/dates very frustrating. It should be a great aid when crossing time zones as a T1D, and yet if you change the time to local time the reader goes into a strop. Users who do a lot of international travel leave the reader at their home time zone. Their mail order system is really bad. When I started using Libre last year the sensors would turn up after a few days, now the website says “up to three weeks” but I find it takes a month.

    The reader needs updating. I have poor eyesight and I would like an ability to zoom into the trends graphs. Since using my Libre my control is very much better (I am 80% within 4 to 8) so I really do not need the graph to have a top value of 25. It should allow me to set that. A lot of the desktop app’s analysis (I’ll come to that) needs insulin and carbs. I find that in a 14 day life of a sensor I will forget maybe 5 or 10 such entries. You only have 15 minutes to remember and “edit” a value. I would like to be able to go through the *user entered data* and be able to edit it on the reader, or failing that, via the desktop app.

    The desktop app needs a lot of improving. It only uses data on the reader, and the reader only keeps 3 months of data, and at that, the oldest month seems to be incomplete. So if you want a record of how your diabetes has improved you have to be diligent in saving a PDF of the data and then compare that data rather than use the desktop app. Any data older than 3 months is lost forever. I would prefer the desktop reader to download the data *to your computer* so that you have a permanent record that you can view later.

    I find the mealtime trends completely pointless. Althougb I have a fairly complete record of insulin and carbs, and I eat at the same time during the day, I find the mealtime trends will rarely use more than 4 or 5 values, most times it says there are no values. And it also does not allow you to set when your mealtimes are. Pointless.

    The creepy. When you connect your reader to your PC the desktop app reads all the data on the reader and uploads it to a US-based website. It does not send your name, but it does send your reader’s serial number. So Abbott has a database with all of your data. Creepy, huh? And since the server is outside of the EU it means that it is *their* data not yours. If you are paranoid about this you can simply disconnect your computer from the internet before connecting the reader to the computer, but that is being a bit paranoid. However, since Abbott has all of your data, why doesn’t it give you access to it?

    1. Tim Post author

      I disagree with everything you say!! Nah, only kidding – you’ve got some really good points, thanks for jotting them down. Here’s a few rejoinders:

      Accuracy – fair point; I think any accuracy issues are completely negated by being able to see trends anyway. It’d doesn’t matter if a single reading is a bit wonky in the context of the 3 billion other readings surrounding it.

      Interstitial – entirely reasonable; but I do fancy some beans on toast now…

      Reliability – yup, I think I’ve had beginner’s bad luck. When you talk about being caught in the rain, was it something like the scene with Mr. Willoughby in Sense and Sensibility? If not, why not?

      Clocks – I’ve only stayed in my own time-zone so far (Edinburgh Standard Time) so I haven’t come across this yet. But it sounds annoying.

      Reader graph – yes, the weird scale that goes up to 25 is really annoying. It gives you the impression that your BG is all over the place.

      LibreView – I think it’s better than reporting tools that have come before it. I downloaded all the data from my pump for the first time in ages and its output was horrible. I totally agree with you, though, that it need improvement. And given the low cost of cloud storage, limiting data to three months is criminal.

      The creepy – I agree on your points about access to your own data. I’m not so bothered about it being exported to the USA as there are data protection carve-outs under the 1998 act and there will probably be similar stuff under the GDPR. Once we leave the EU (if, indeed, we do) then that is another matter…

  2. Antony


    Thank you for writing this post and for those that have commented. I have just started on the FreeStyle Libre, so will bare all this in mind.



Leave a Reply

Your email address will not be published. Required fields are marked *