Sunday, February 19, 2006

Don't pull out that checkbook just yet

I haven’t been updating very often. I kept hoping the initial issues would be straightened out and I’d be able to tell you how wonderful the integrated pump and continuous BG monitor was. I knew it wouldn’t be perfect, but I had hoped it would at least be useful. I’m not so sure it is. Here’s my opinions, in bullet form:

What they got right

The display. One click on the pump and I see a graph of the previous 3 hours, plus my current BG reading, plus an indicator if things are changing “rapidly”: one arrow for rising/dropping a point a minute for the last 20, 2 arrows if two points a minute. This screen is probably the most useful thing about the whole setup.

The display part 2. There is a second screen which shows you the previous 24 hours, same extra info, which is really great when you wake up and want to see what the heck happened overnight, or is my bizarre overreaction to breakfast today as bad as it was yesterday, etc.

The software. As part of the study, you download the pump to a website, where you can create various reports of the data. My favorite is the one where you can overlay the past X days of sensor readings, to allow you to look for patterns. Sure, there are improvements that could be made to the interface for this, and more flexibility would be nice, but in general it’s a useful tool the way it is. Which brings us to –

What needs more work

Accuracy. It’s hard to decide where to start on this list, but I guess this is the biggest issue. I mean, we all know that the BG reading you see on your meter is pretty much an imaginary number. Try it twice in a row, or the same drop of blood and strips in your meter and back up meter, and you’ll see ten to fifteen points difference most of the time. Plus, sensor is measuring BG in interstitial fluid, not blood. So I know I shouldn’t be upset if the sensor and the meter disagree by 20 points or so. The problem is, they differ by much more than that an awful lot of the time. You pretty much have to ignore the first twelve hours or more, where I’ve seen the sensor and the meter as much as 100 points apart. (See warm-up time, below). But even when you’re on day 2 or 3 of the sensor, sometimes the readings are just out to lunch. Like 190 versus 286 as a post meal reading.

The BG meter. OK, this is a study thing, but maybe it’s contributing to the accuracy mentioned above. The deal is, all us study participants must use the same meter and strips. Makes sense, doesn’t it? Except, there’s a problem with the meter we’re using. It’s not great below 100. %$#@&! Hello? Below 100 is when you NEED the meter to be accurate. Here’s the deal. I tend not to feel a low until I’m about 55 or so. I’ve been working at avoiding them (I averaged one a day below 55 before the study started) and now only seem to get shaky every other day or so. But, when I check with the meter, I was seeing things like 94, or 78. Hmm, guess I was wrong. Well, then I get a call from our study coordinator saying that they think there is a problem with this meter in the normal-to-low range, so if you feel low, treat it anyway, or double-check on your old meter. So the next day I felt weird; the sensor showed me happily buzzing along at 110 or so, and I decided to do a comparison. Big drop of blood, half on old meter (fresh bottle of strips), half on the meter for the study, the one the sensor calibrates against. The result? Study meter says I’m 78. Old meter says 53. I feel 53. Making the sensor exactly DOUBLE my best guess as to my actual BG.

Sensitivity. I try to look at the picture more than the number. Hey, the numbers may be off, but at least it shows you whether you’re climbing or falling. Except when it doesn’t. Take this morning for instance. When I went to bed, sensor thought I was 156; meter said I was 136. No big deal – in the ballpark, if slightly lower than I like to be at bedtime. I decided to leave it alone and see how I did overnight. I woke up, and the display screen showed me relatively flat all night and currently 142. I checked with the meter and got 85. WTF? I dropped 51 points overnight, not 14. There’s a big difference there. To be fair, I realize this is a complex thing to measure, the sensor degrades as your body tries to “coat” the thing under your skin, which is why you have to keep re-calibrating it with the meter. BUT, as a consumer I’m just saying that I don’t like it.

Warm up time. You have to change the sensor every three days. I do it with my site change. After you put the sensor in, you tell the pump about it, and it goes into a countdown for 3 hours, and then beeps to ask you for its first calibration reading. About 15 minutes after the calibration your first reading shows up on the pump. So far, so good. The problem is, that it takes a lot longer than that to actually be accurate – or as accurate as it gets. Some times it seems ok in the afternoon after my morning site change, some times it’s not synched up until the following morning. Some times it never syncs up. Then, at 36 hours exactly, it beeps and shuts off. The reasoning (I assume) is, that it degrades over time (hah!) and they don’t want you using the meter values as gospel when they know it’s off. As opposed to the rest of the time, when the inaccurate values are apparently OK. It would be better if at 36 hours it gave you a warning (like a low reservoir warning) and then shut down in another 6 or so.

Quality control. The first box of sensors I got, half of them went bad. Most of those were during the warm-up period, before I ever got a reading from them. My study coordinator actually replaced the little transmitter too, in the thought that it might be that rather than the sensors going bad. Afterward, in the second box, I’ve only had one go bad. BUT, keep in mind in the real world, I’d be paying for these. Let’s assume these sensors sell for the same price Wil pays, about $40 a pop. I’m ignoring the cost of the transmitter and the pump itself here. That means I tossed about $200 without ever getting any benefit from them. Add to that the warm-up day, the accuracy problems, etc, and it comes out to about $40 a day when you actually get useful readings from them. Would I use it if my insurance paid for it? Yeah, in a heartbeat, because it’s at least better than what’s out there right now. Would I use it if I had to pay $15 a day on my own? Probably. We can afford that, if we’re careful. Would I use it as it stands now, where I average 1 day in 3 of useful information – $40 a day for the days that it works? No way in hell.

Transmitter. The transmitter which sends the readings to the pump is not small. It’s an oval shape, the length and thickness of my pinky. You stick it to your skin with an adhesive pad that starts to itch after about two days and leaves behind a sticky residue That you have to really scrub at it to get off. A lot of times, it doesn’t deal with the fact that while I’m sleeping the pump might be on the other side of my body. It beeps politely to let me know there’s a weak signal, which I sleep through, and eventually shuts off sensor communication. At this point the pump beeps again, and again, until it finally goes into beep and vibrate mode and wakes me up. I turn on the light, clear the alarm, and re-start the sensor, which means that in 15 minutes it beeps again wanting a calibration, so I can’t go back to sleep until that happens. Now on its own this wouldn’t be a big deal. In fact, if this was the only problem with the whole setup I wouldn’t even mention it, but it is an irritation. And the fact that it is this thing that wakes me up, rather than a low alarm when I’ve obviously gone low overnight, this just pisses me off. I’ve had to reset the low threshold to alarm if I dip below 100 to have a chance of catching it before the real BG hits 50. What I think the issue is here is that they are using the technology because it’s cool, not because it’s useful. Yes, it’s really cool that the sensor/transmitter broadcasts a (random) BG reading to the pump, and that the pump displays them for you. It’s great that you can do this without wires. BUT, what the developers seemed to have forgotten, is that we’re already hooked up to the frigging pump. Why not just run a wire along the infusion set tubing, with say 12-18 inches of free wire at the end so you can position the sensor away from your infusion site. That way, you’d definitely change it with your site, AND you wouldn’t have to worry about this lost sensor crap. Plus I wouldn’t find the damn thing alarming every time I step out of the shower, because it would KNOW I was disconnected.

OK, I’m not going to go into the minor annoyances at all, because if the major ones were fixed I could deal with the minor ones. My husband is amazed I’ve stuck with the study, but I want them to see what’s wrong, so they know what they should do before they make this a commercial product. I mean, it COULD be so damn good, if it only worked. In the meantime... Do you remember all the hype before Windows 95 came out? How great it looked, how much it did? And then when you actually used the real product, before the service packs came out? This is worse. This is Microsoft Bob (or the Paper Clip Guy, for those of you fortunate enought to have never seen Bob). It’s something that in theory could be really useful, even indispensable, but in reality it’s just a giant pain in the butt. I really hope they can get this to work – or someone else can – but I don’t think I’m going to be the first to jump on the boat to start throwing money at a 1.0 product.