Abstract
This study evaluated a next-generation automated insulin delivery (AID) algorithm for Omnipod in type 1 and type 2 diabetes across multiple phases: 14-day run-in with usual therapy, 48-h AID use in a hotel setting (type 1 only), and up to 6 weeks of outpatient AID use. Participants did, or did not, deliver manual boluses at alternating periods. Twelve adults with type 1 diabetes completed the hotel phase; 9 of those 12 plus 8 adults with type 2 diabetes completed the subsequent outpatient phase. Outpatient % continuous glucose monitor readings >250 mg/dL decreased from 33.5% at baseline to 9.4% with, and 14.3% without, manual boluses in type 1 diabetes and from 20.8% to 7.7% with, and 10.5% without, manual boluses in type 2 diabetes. Time below 70 mg/dL remained <4% during all phases. No adverse events occurred. In conclusion, a next-generation AID algorithm demonstrated feasibility in people with diabetes.