Background: MiniMed™ 670G system user-dependent settings influence glycemic outcomes in persons with type 1 diabetes. The ability of a mathematical model-based digital-twin program to personalize and automate user-dependent MiniMed™ 670G settings was assessed.

Method: A small feasibility study recruited experienced MiniMed™ 670G system users (N=19 users, >14yrs) to complete an at-home 3-weeks run-in phase with Auto Mode enabled followed by a 2-weeks study phase. Run-in data were used to generate a digital twin that personalized system settings during the study phase that comprised a non-randomized Auto Mode and Manual Mode arm. System use and glycemic outcomes were assessed.

Results: Personalized Auto Mode (Auto Mode+) reduced the mean(SD) of sensor glucose (SG) from 151.1(12.3)mg/dL to 149.3(11.6)mg/dL and increased time spent in target SG range from 74.0% to 75.9% (Table). Personalized Manual Mode (Manual Mode+) demonstrated a mean SG and TIR of 150.1mg/dL and 67.4%, respectively.

Conclusions: Improved Auto Mode therapy and acceptable Manual Mode (sensor-augmented pump) therapy were observed utilizing a digital-twin program to personalize MiniMed™ 670G system settings in subjects whose glycemic control was excellent at baseline. This is the first demonstration that a digital-twin can successfully automate and personalize insulin pump therapy.


B. Grosman: Employee; Self; Medtronic. A. Roy: Employee; Self; Medtronic. D. Wu: None. N. Parikh: None. L.J. Lintereur: Employee; Self; Medtronic. N. Schneider: None. R.L. Brazg: None. S.K. Garg: Advisory Panel; Self; Boehringer Ingelheim Pharmaceuticals, Inc., Eli Lilly and Company, Medtronic, Novo Nordisk A/S, Roche Diagnostics France. Research Support; Self; Dexcom, Inc., Eli Lilly and Company, Medtronic. R. Vigersky: Employee; Self; Medtronic.

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at