Anger and fear after popular diabetes app breaks
bbc.co.ukWearing a Libre 3 sensor from Abbott, I can say I'm happy I don't need the official app. The sensors are good. The app however forces you to be always online and send your measurements to Abbott's servers, if you want the measurements to be available to other apps or devices. It means that if you lose connectivity, or Abbott loses their servers, no integration will work anymore. Your smartwach will not display your readings, your partner will not be able to see them on their device, and in my case, if I used their API, my closed loop, automated setup for insulin delivery would stop regulating.
How they arrived at that design decision I can only wonder. Probably not even malicious, just taking the easiest path. And now they even managed to get their app dropped from Apple's app store. I understand people don't want to tinker with their setup, but relying on closed systems like Apple's and Abbott's risks exactly such scenarios.
I'd like to kick Abbott entirely. And I keep trying to switch to competitor Dexcom, who are a little bit more open. But their hardware just doesn't work as well for me.
Watch out the Dexcom has unmutable alerts, including a rigid 6 hours to go on your sensor, even at 4AM !! It blows up, "F*&%k You, our getting the next sensor into your arm right on time is more important than you getting to sleep" alarm blares. So just watch out for that "feature". I'm now a Libre 2 user. I tried the first one and they can get knocked off your arm very easy. I then tried the Dexcom system and I liked that the reader was separate but had a cool feature that it could also be pared to the phone. What really drove me off the Dexcom is that it is a 10 day sensor, so you end up with having to change it different days of the week. And it wasn't as accurate the first day or so as you calibrate it. Finally the app and the software in the portable reader were insane. The dexcom was willing to blow up and make noisy unmutable alerts, for awhile I was sealing the reader in a cigarette case to muffle it. The final straw is that it treats the impending expiry of a sensor as a catastrophe. It actually woke me up at 4 am with the notification, which can't be disabled that the sensor was going to expire in 6 hours. So back to the freestyle where they treat you like a criminal if the sensor gets knocked off, or if the applicator fails. You have to have the serial number ready and they make the process to get a replacement very stressful. I noticed that the Freestyle instructions only allow you to pair the sensor with one device. I'm glad I really read the literature and decided to get their reader, which also has a traditional blood glucose strip reader. And the thing is passive I have to actually scan my sensor to get a reading and a graph of the recent trend. So the Freestyle is nice overall but I had to invest time to get a good life balance with it. I feel bad for the average crowd without tech or science literacy.
I wouldn't get the alarm because Dexcom works fine with Xdrip. But how I wish I'd get to day ten!
So far none of the seven Dexcom G7 sensors I tried made it to ten days. Three came off within one day, one was a dud, one failed from the start, and two made it to around day six before becoming unreliable and failing eventually.
My experience last year with G6 was that they'd start getting unreliable after day six. At least they stayed on :-/
So maybe I react with inflammation or something to their sensor probe. They just don't work for me. And the glue in the G7 doesn't stick either, which is a new problem.
So Abbott it is, luckily without their patronizing app.
Have you tried getting those bandaid stickers to better secure them? One of the two companies has them for free, but you have to ask. They are on the usual sites.
I want to know how can they get such slipshod systems past regulators. The design you have described sounds terrifyingly unsuitable for a medical device.
The logic is sorta backwards. They can’t fix anything because it’s such a pain to get it re-past regulators. Making a large software change would be huge excel documents scrutinizing the FMEA and “risk assessments”, and if any risk increases or decreases, goto 510(k)
My assumption is that it was easier to get past regulation that way. Because their sensor is only talking to their app. And the app is only talking to their server. So no third parties to interface with which would complicate things.
And what happens on the server probably doesn't need certification.
They do however go out of their way to make it hard for third parties. They encrypt their app and run integrity checks to detect patches. Now writing this again makes me feel like they purposefully block third parties, not just ignore them.