Can some please explain the closed loop in easy to understand words? In research, I’m getting a lot of peer reviewed tomes with more info than I need.
Thanks.
The pump looks at your CGM and adjusts insulin accordingly. It does what you do when you are not sleeping. I therefore think that the biggest benefit of a closed loop (or AP=Artificial Pancreas) is during the night, not during the day.
So the pump makes insulin decision? Based on the CGM?
Yes.
The artificial pancreas systems that I have some limited familiarity with (OpenAPS and the Loop) use CGM data to set temp pump basal rates (plus or minus) based on a program prediction for near term future glucose levels. Each change of insulin dose via temp basals is only set for five minutes at a time and the program will assess blood glucose every five minutes and make a new decision. It is a conservative system designed to “fail” to the pre-set basal program settings. It’s like having an automatic dynamic basal rate. These systems do not bolus dose; that is still up to the pump user.
2 Likes
No, whatever algorithm that you are using controls the pump based on the information from the CGM.
The CGM sends its data into the cloud (or directly to the controller).
Based on the algorithm, the controller (in DYI circles this is a smart phone) sends instructions to the pump.
The Pump then administers the basal dose appropriate (or the correction if you’re asleep).
CGM→ Cloud→ Controller→ Pump
Not that it matters much in this thread, but to be clear Cloud is not necessary. In most cases, DYI systems work as follows: (CGM and Pump)→ Controller→ Pump, although the version with Cloud for CGM data in particular is also a possibility. To make a decision, the controller algorithm needs to know both BG data from CGM, and pump data (e.g. bolus) from the pump.
1 Like
Mind you, I’m not tech-savvy enough to want to do that. The limiting factor for me is that I don’t have extra pumps sitting around that can receive instructions from a controller.
1 Like
Which system do you use? I have built and used both OpenAPS, and Loop, and I know both can and do work without sending data to the cloud. We do send everything to the Cloud (Nightscout) for visualization, but this is not necessary for operation of the control algorithm. In some cases, people do get CGM data from share or Nightscout to the control algorithm, but that’s just an option.
There are several parameters involved in the algorithm. Carb intake, bolus, insulin onboard, CGM and pump.
Why carb intake? Are there algorithms that are determining and administering bolus insulin doses?
There are Bolus calculators in Loop which propose a bolus. And carbs + Bolus (as part of onboard insulin) are needed for forecasts, i.e. immediate or future temp basal adjustments.
My pump already has a bolus cackalator, so it wouldn’t occur to me to have two “advisors.”
The whole point is that you don’t touch your pump manually anymore. If you have functioning bolus calc parameters in the pump, then you just copy them into the phone app and use the app to calculate the bolus and LET the app control the basal.
Also, if you don’t enter bolus, the pump will maximize your temp basal in order to compensate for the missing bolus, but of course, it will take at least 30 min instead of immediate injection of x units.