It’s yet again Diabetes Awareness Week and we’ve just had ADA. It’s a perfect time for Awareness of Diabetes as well.
But when we talk about Diabetes Awareness, it shouldn’t just be the good stuff. There’s also a lot of not so fun stuff that goes on. Whether that’s at an individual level (Burnout, DKA, etc) or at a societal level (Stigma).
It’s also a time when people talk about how great Diabetes Technology is, and how it’s helping to make life easier. And that’s true, but we shouldn’t let it give the wrong impression.
As an advocate for the use of technology in Diabetes of all types, whether that’s greater access to various testing technologies, smart pens, pumps or automated insulin delivery, and an active member of the #WeAreNotWaiting community, I’m all for making life easier with tech.
But tech has a down side. It goes wrong. And when that happens, it can go wrong spectacularly. This Diabetes Awareness Week has been #BadDiabetesWeek2022 for me for precisely these reasons.
Bad Diabetes Week 2022?
Technology has two major concern areas. The first is the tech itself, that can go wrong, or fail, but its usually fairly obvious when it does and can be quickly dealt with. The second is the human/tech interface. This one can be much trickier, with the effects not being obvious for some time, and the alerts tonit not being immediate. Let me give you two examples…
The failed cannula
This one, on Monday, saw my cannula bend, and not be able to deliver insulin. The cannula was changed at 10am, and then a nice “healthy” lunch of Katsu Curry and Donuts followed. I’d normally expect a bit of a spike as a result of that meal, but not what you can see above.
So I felt around a bit. As I use the Roche Insight pump, I use the accompanying cannulas. These have a piece that is hinged and that piece had popped up and yanked the cannula out. I can only assume that I had caught it briefly and hadn’t noticed. That, combined with the use of u200 insulin, meant that the cannula wasn’t especially damp, so I hadn’t noticed.
Fail number one for the human/tech interface.
The occluded tube
The second occasion in Bad Diabetes Week 2022 came after a shower. I’d taken the pump off for a shower and suspended the insulin, then reattached and restarted after it.
Around three hours later, my pump alarmed to tell me about an occlusion, as the glucose levels climbed. Just as I was about to get on to a packed tube train. At this point I got on the tube and went to work, not fancying a reset in the busy station. The result? See below…
The area in Blue is the insulin delivery I suspect I either didn’t get or got very partially.
When disconnected the cannula and attempted to fill the tube, nothing made it through. Fully blocked. A quick tube change and then we’re back to insulin delivery as normal, and glucose levels coming down. Slightly annoying as for me, the levels were too high to comfortably have breakfast, but nothing serious happened.
Okay, you had a couple of tech issues, but so what?
That’s a very good question to ask. The key takeaway here is that AID systems, even those that have multiple hormones and allow fully closed loop with no mealtime or exercise related actions are not “intervention-free”. And we need to be careful when talking about them as such. Mechanical failures between ourselves and our technology will always happen. There’s no way around it yet.
But that’s important. While we have tools that will alert us to ever increasing glucose levels, and perhaps pre-empt some of the possible issues, we have to be careful about casting AID systems as a panacea. The issues identified here still happen, and anyone using CGM and a pump experiences them in the same way as an AID user. If they don’t react well to these types of issues in that scenario, would the fact they were using an AID make any difference?
Similarly, day to day activities that are linked to the means of insulin delivery still need to be done. Set changes. Pod changes. Reservoir changes. Tube changes. Sensor changes. Maybe even sensor calibration. Without all of these, the technology works far less effectively, if at all.
That’s why, when we talk about automated insulin delivery, it’s important to remember that it’s not just an algorithm. There’s a whole load of other stuff that requires improvement, and we’re starting to see that. 7-day sets and calibration free sensors are a start. Pre-filled cartridges make sense but are far from widespread. Imagine if a pump could just use a penfill cartridge! It’s not just the failures…
All the automation in the world is only as good as the interventions of the user with the human/technology interface. Limiting those is a key part of obtaining better outcomes.