Just flagging up that Medtronic Canada have put up a handful of videos for Healthcare Professionals. This one is the best of the bunch in my mind (but my French isn't so good, so Nouveaux developments concernant le pancreas artificial might well be better :-) )
The video is almost 2 hours long, so don't expect a Twitter / quick file take home messages :-)
There's a whole range of useful and (to my mind) interesting information here. Some of it I've heard before of course, but it does no harm to hear messages reinforced*.
(* so long as there's actually some good data driving them of course...)
I've jotted down some quick examples from the video, in case you want to skip through the video:-
Insertion (~30 minutes)
As some of us have heard before, don't press down hard on the inserter
Don't tape the transmitter tightly with the tape flap, but do press down on the adhesive patch once applied (including the sensor hub area).
They also talk about mounting the sensor vertically (which makes no discernible difference for us, but great if it helps some people).
They also mention they have a Sensor Performance Report (SPR) that gives them useful information, flagging, for example, calibration anomalies (e.g. plugging in an old BG...). They (Medtronic) have this. We (the users) don't. Why Medtronic? Open data please...
MARD (~55 minutes)
The Damiano paper...
This one obviously hit a nerve inside Medtronic HQ.
Essentially, they suggest some data should have been excluded from the paper's analysis (ie if you were wearing the pump it would have alarmed to say ‘something is wrong’ and you wouldn't have used the data). Medtronic post processed per intended use (i.e. excluding that data) and calculated an Enlite MARD of 13.89%, not 17.85% as per study, which is close to the 530G User Guide’s MARD of 13.63%. They also question the number of sensors (n=23) used in the study, implying, I think, that there is insufficient statistical power in their data.
“Inaccurate information has been provided” by this study - ouch!
Fortunately, I've not been in a similar position with any of my published papers in my area of work, but I am puzzled that the Editors (and there are several versions of this / very similar paper out there) haven't asked the team for a response to Medtronic's concerns. Maybe Medtronic haven't formally written to the editors (that would be standard practice to raise a concern or question over a peer-reviewed paper)?
By around 65 minutes in, there's a comparison of hypo-region MARD and Medtronic's ability to predict possible hypos thirty minutes out (essentially what SmartGuard looks out for on the 640g).
Low Glucose Suspend (~70 minutes)
They have a huge data set via CareLink - around 3 million days of data. I knew CareLink was big, but that really is big data...
They clearly demonstrate fewer hypo events, but does low glucose suspend (as on the Veo / 540g) cause more hyperglycaemic episodes? I'll leave you to decide (from their data anyway), but I was puzzled at their use of 16.6 mmol/l as they trigger? Surely you'd want to use ketone production limit (13.9 mmol/l) or even a sensible upper "target" range value...
2/3rd of Low Glucose Suspend events happen at night time.
At around 90 minutes in, there's clearly an attempt to give the positive message that for many people, if you shut off insulin, the body does a good job of turning round a hypo. There's more than half a suggestion perhaps not to give carbs whilst Low Glucose Suspend is active: "if you don't take carbs you're fine"... I appreciate this is clearly aimed at Healthcare Professionals and - hopefully - they're going to think things through, but I had a little shout at the screen at that point... Looking at an average (or even a single standard deviation range) on big data does not in my humble opinion empower a statement like that. You, your patient, or loved one, might not be within that 90% (or whatever) that don't (usually) need carbs. We certainly have a more sophisticated "algorithm" in our heads of what to do when. Big data does not give one 'fits all' answer.
Sensor innovation (~95 minutes)
Electrochemical Impedance Spectroscopy (EIS)-based diagnostic chip monitoring sensor health, dual active sensor areas (e.g. giving ability to take data from both or just one sensor area if an error is spotted) and more positive news. We'd certainly welcome anything that improves the less than perfect performance we have from our current batches of Enhanced Enlite Sensors.
Thanks to Leo for spotting these videos had gone online!
No comments:
Post a Comment