An "integrated" pump/CGM? Medtronic doesn't fit the definition.
I often see comparisons between the Minimed/Medtronic CGM and the Dexcom, and often one of the perks of the MM is that it is "integrated" with the pump.
From my perspective, it is not. It shares a display and a housing, but nothing about the two systems are truly integrated. No data is shared between the two in order to give a useful alert. (I'm not talking about a closed-loop system, that's an entirely different, more complex, ballgame altogether).
What do I mean by that? Here are two examples:
1) For the "pump" side of the device, I set my high and low targets for the bolus wizard to figure out how much insulin to administer. On the CGM side, I also set, separately, high and low targets (the menu sequence is even different). Why? Well, I need to do that - in part because of #2.
2) Looking specifically at the "Predicted High" alerts, the amount of Active Insulin (Insulin-on-Board, IOB) is never considered. This has a huge ripple-effect.
During daytime hours, I have my Bolus Wizard set to a target of 90-110 mg/dL. My Sensor (CGM) target is 80-175. I don't particularly like the 175 target, and in fact my upper limit is lower overnight (since I have no postprandial events while sleeping), but it's the only way to avoid getting yelled at by my pump when my BG goes up slightly after each meal.
Now, the following two very real scenarios show why I say the two devices are NOT integrated:
A) If, after a meal, my BG is rising, I get a "Predicted High" alert, even though I have plenty of IOB to cover it and bring it down. (Ideally, it never actually reaches the High mark). To avoid these superfluous alerts, my High threshold (on the CGM) is higher than I'd like it to be.
B) If, after the postprandial spike is over, I'm hovering at 130-140, I would want to know about it. If there's insulin on board, fine; but with no IOB, I'd like to be alerted that a small correction might be in order to bring me to my target of 110. This could do wonders for lowering my A1C, as going hours at a time with a slightly elevated BG and no IOB should be easily correctable.
Please forgive the rant, I just get frustrated when I hear that the two devices are integrated, when in fact they are not. Such a simple programming change would be a huge help. Anyone know how to get the message across to the developers? Any other thoughts?
Replies to This Discussion
-
Permalink Reply by miketosh on -
There are significant risks of adding that much integration, which is why MM and others prefer to call those "Artificial Pancreases", to give them a new category of device.
As a side suggestion, I've found that timing of my bolus affects how much of a spike I get after eating. In general, for high carb meals, a 10 minute delay after bolus before eating is needed. For low carb meals, I can bolus as I start to eat. No matter what I eat, if I bolus after that first bite, my sugar will soar into the upper 200's or higher.
-
-
Permalink Reply by Scott E on -
The Closed-Loop/Artificial Pancreas is meant to dispense insulin automatically. I'm only referring to the timing of alerts. Is there really a risk to that? (The alternative would be to lower my sensor target significantly, to the point where I grow numb to the nuisance alerts and ignore all of them... is that better?)
-
-
Permalink Reply by Capin101 on -
I do not think anyone argues that CGMs are a perfect device, but many do find them another helpful tool. There are certainly areas for improvement.
The only area I may I may be able to provide you some advice on are your predictive alerts. In Sensor -> Edit settings you have a Predictive alert feature. I have played with these numbers and decided that 20 minutes works best for me. I suggest you may want to experiment with them. If you set the predictive alert too low (5 minutes), then you will already be at your upper or lower limit when the predictive alert is activated. If you set your predictive alert too high, then you will get a lot of false positives (like your PP spike).
I personally like to be notified when my PP spike is kicking in (if its on the high side) as I can be more vigulant at that time to see if I stay in my targets, but on the other hand having Forrest Pump vibrate at me every 15-20 minutes can annoy me like no other!
-
-
Permalink Reply by jrtpup on -
Scott, I'd love to see real integration/communication between my Ping & Dexcom. It doesn't seem, from a programming perspective, as if it would be too difficult, except for the FDA piece.
When my dex wants a calibration, I use my pump remote/meter to FS, then have to enter it in the dex. Would be nice if it could enter itself. I keep my low/high on my CGM at 70-120 24/7, but my target on my pump is set for 90.
Yes, the IOB piece would be wonderful. As you said, the only way to tell if there's IOB when I get a high warning is to FS. It would be really nice to skip some of those many FS, and I find that those are the ones I sometimes resent. High warning, check BS, pump calculates correction bolus and says 0 because there's IOB. Waste of a FS and time.
As for getting the message across, I mention it to the Animas and Dexcom folk whenever I have to call. I'm sure someone will have a better way ;)
-
-
Permalink Reply by Natalie ._c- on -
Well, Medtronic DOES read this board, and so I'm sure it will come to someone's attention. Whether they will think it is doable/worth the investment is another question.
Since I still have a 722, the predictive alert stuff is a mystery to me -- my question about that is what if you're going up rapidly and it alerts, but then you flatten out and don't actually hit your designated high point? Is that just another annoying alarm then?
Personally, I've used MM for almost 13 years, and never been unhappy, but I REALLY wish I had a carb-counting library, because I'm REALLY lousy at it. I wonder what it would take to add that to the next generation of MM pumps.
-
-
Permalink Reply by Scott E on
-
There isn't a "Just Kidding", or "Sorry, we Goofed" alert (though that would be nice!). You just get the High Predicted and that's it.
-
-
Permalink Reply by Duarte on -
Hi Scoot
there is some integration between sensor and pump if you go to low the pump suspend automaticaly you basal .
At least the european version does that . -
-
Permalink Reply by Scott E on
-
Duarte, you are correct. I neglected to consider the LGS feature that is available outside of the United States.
While LGS does indicate a rudimentary connection between pump and CGM, I'm not so sure I would call it integrated, in the sense that I described above. It doesn't improve the information given to the wearer, nor is information from both subsystems (CGM reading, IOB calculation) used together to generate an action (alert). It's only a CGM reading which triggers the act (suspend delivery).
So while one does operate the other, in a sense, it doesn't affect the way a person would use either device. It's an emergency shut-off, something that the wearer would likely never see happen (either because it doesn't, or because he's not in a conscious enough state to witness it).
-
-
Permalink Reply by Duarte on
-
Your correct it isnt a fully integrated system ,but if someone built one with the actual accuracy of the sensors will you trust in such system to manage your diabetes ?
I use the Lgs fuction in my pump and it is setup to shut the basal at 60 it shut of my pump at night several times .if the pump doesnt have shut off probably i wull not be here commentig .So i prefer pour integration that no integration.
-
-
Permalink Reply by MegaMinx on -
I used MM CGMS for 3+ years, and now use Dexcom. But both MM and Dexcom clearly state that 'corrective' action (to treat high or low) should ALWAYS be based on a meter BG. I've had my share of 'false' and annoying alarms. But also many that have correctly alerted me.
But if I can't trust the number on the CGMS, I certainly don't want the pump to take it into account for calculations, or decision to alarm. I do think that once accuracy is improved, there will be more 'talking' between them. I think you have listed some great examples that would be beneficial.
-
-
Permalink Reply by Scott E on
-
Of course the decision to take corrective action is supposed to be made by fingerstick data, not CGM data, but I see no harm in sounding an alarm because of CGM data -- after all, that's what the CGM is for.
But to me, an acceptable postprandial high is higher than a high at other times (does that make sense?). All the information to know if the CGM reading is right after a bolus is right there in the pump, so it would be helpful for alarms to sound based on that information. Again, not to make a therapeutic decision, but just to bring something to attention.
-
Advertisements
Groups
TuDiabetes Team
DHF STAFF
 Manny Hernandez(Co-Founder, Editor, Patient) |
 Andreina Davila(Co-Founder, Patient Spouse) |
 Emily Coles(Program Manager, Patient) |
 Emily Walton(Office/Volunteer Coordinator) |
DHF VOLUNTEERS
Lead Administrator
 MelissaBL |
Administrators
 Bradford |
 Gerri |
 Lorraine |
 Marie B |
 Teena |
Spread the word
This site complies with the HONcode standard for trustworthy health information: verify here.
© 2012 A community of people touched by diabetes, run by the Diabetes Hands Foundation.
