Algorithm-driven insulin dosing
A University of Virginia study found algorithm‑guided insulin dosing paired with continuous glucose monitoring improved blood‑sugar control versus standard titration — the algorithm analyzes two weeks of CGM data and issues weekly insulin adjustment recommendations. Users reportedly accepted the automated guidance at high rates, suggesting practical adoption is realistic in type 2 diabetes care. (news-medical.net)
The trial was published in Diabetes Technology & Therapeutics (DOI 10.1177/15209156261420193) with Anas El Fathi and Marc D. Breton listed among the lead authors. (journals.sagepub.com) Thirty adults were randomized 2:1 in a two‑site, 16‑week study that enrolled CGM‑naïve participants with baseline HbA1c between 7% and 9%, and all participants used insulin degludec during the trial. (journals.sagepub.com) The predefined primary endpoint was change in CGM-measured time in range (70–180 mg/dL); the experimental arm gained a 20.3 percentage‑point increase versus an 8.3 percentage‑point increase in the control arm, yielding an estimated treatment difference of 14.6 percentage points (one‑sided 95% CI lower bound 4.0; P = 0.005; exploratory two‑sided superiority P = 0.03). (journals.sagepub.com) Hypoglycemia was uncommon: median time below 70 mg/dL was 0.34% in the experimental group versus 0.00% in control, level‑2 hypoglycemia rates were about 1.1 versus 2.2 events per patient‑year, and the study reported no severe hypoglycemia or serious adverse events. (journals.sagepub.com) The published report states CGM‑based dosing recommendations were “well accepted” by study physicians, and the university press materials additionally describe strong participant acceptance without listing a numeric acceptance rate in the paper itself. (journals.sagepub.com) The authors and trial registry (NCT06111508) note the intervention used a Dexcom G6 CGM and that investigators called for confirmation of long‑term impact in broader populations through larger, longer trials. (clinicaltrials.gov)
