Hey readers! 👋
This week, algorithms are doing the heavy lifting. From CGM-powered insulin dosing that's dramatically improving time in range, to machine learning models predicting T1D progression before diagnosis, the software side of diabetes tech is having a moment. Whether it's closed-loop systems helping kids just be kids or AI-driven titration outperforming standard care, the theme is clear: smart algorithms paired with good data are reshaping how we manage diabetes. Let's dig in.
🧠 This Week's Highlights
New CDT Algorithm Can Better Manage Type 2 Diabetes, Study Finds - The University of Virginia's Center for Diabetes Technology tested an algorithm that uses two weeks of CGM data to generate personalized weekly insulin-dose recommendations for people with type 2 diabetes. In a 16-week randomized trial, participants following the algorithm's guidance spent 75.3% of their time in range, up from 54.1% at baseline, while those who self-monitored improved only modestly from 50.2% to 55.3%. That's a striking difference. – Center for Diabetes Technology
"These results clearly show that diabetes technology and advanced algorithms can be leveraged to great effects, well beyond the classical paradigm of automated insulin delivery."
What's particularly interesting here is the acceptance factor. Clinicians noted how well participants embraced the technology, suggesting that algorithm-driven titration isn't just effective, it's practical for real-world adoption. Larger studies are needed, but this is a promising signal for expanding CGM-algorithm pairings beyond automated insulin delivery.
Tandem Reports Positive Findings for Control-IQ in Low-Bolusing Youths - In a study of 98 adolescents who rarely use manual bolus doses, Tandem's Control-IQ system delivered 16% more time in automated mode (about four extra hours daily) and 8% higher time in range compared with a patch-based AID system lacking automatic bolus capability. – Drug Delivery Business
"When bolus frequency is low, AID systems with automatic bolus delivery can make a meaningful difference and should be considered when choosing an AID system."
This matters because low-bolusing youth represent one of the trickiest populations to manage. An algorithm that compensates for missed or infrequent boluses can meaningfully close the gap, and that's exactly what Control-IQ's automatic correction bolus feature is designed to do.
📊 CGM Algorithms Expanding Their Reach
Dexcom Highlights Benefits of CGM for Type 2 Diabetes at ATTD 2026 - At ATTD 2026, Dexcom showcased Smart Basal, an AI-driven basal insulin dosing optimizer that lowered average glucose by roughly 40 points and increased time in range by 20%. The company also previewed its expanding digital ecosystem, including Stelo AI food logging and plans for multi-analyte sensors targeting broader populations. – Drug Delivery Business
"People wore CGM at the beginning, before they used Smart Basal. They used it for about a month, and then we followed them afterwards, and we saw average glucose drop about 40 points."
The expansion of CGM algorithms into type 2 diabetes and even prediabetes populations signals a significant shift. These aren't just monitoring tools anymore; they're becoming intelligent decision-support platforms.
How Technology Is Changing Type 1 Diabetes - Texas Children's Hospital details how closed-loop AID systems are transforming pediatric T1D care, improving A1c and time in range while reducing daily management burden. The center is also piloting remote monitoring and machine-learning tools to make CGM data actionable for clinicians. – Texas Children's
"What I notice in my practice is the burden of diabetes management being lifted by these systems, allowing my patients to think less about diabetes and more about just being a kid."
Their advocacy work is equally notable: securing full Texas Medicaid coverage for CGM in 2020 and now pushing for closed-loop system coverage. Access remains the bottleneck, even when the algorithms work.
🔬 Research & Prediction Models
Predicting the Metabolic Inflection Point in T1D Progression Using Machine Learning - Researchers developed machine-learning models that predict the metabolic inflection point preceding clinical T1D in autoantibody-positive individuals, achieving an AUC of 0.77 at 1.4 years before diagnosis using OGTT-derived features. – American Diabetes Association
This is a different kind of algorithm, one focused on prediction rather than management. Identifying who is approaching clinical T1D with over a year's lead time could reshape how we think about early intervention and monitoring in at-risk populations.
Comparison of Open-Source Loop and Tandem Control-IQ in Adults with T1D - A new comparative study examines glucose management outcomes between the open-source Loop AID system and Tandem's commercial Control-IQ in adults with T1D. Full results aren't yet available, but this head-to-head comparison is one the community has been waiting for. – PubMed
DAYTime Trial: CGM for Ugandan Youth with T1D - A five-year RCT will evaluate whether CGM improves time in range and is cost-effective for 180 Ugandan youth with T1D, a critical study for informing management guidelines in low-resource settings. – PubMed
📱 Quick Bites
Best Type 1 Diabetes Apps of 2026 - T1D Strong's roundup of the top ten apps for CGM data sharing, analytics, gamification, and mental health support. Worth browsing if you're looking to refresh your toolkit. – T1D Strong
NICHE Cell-Therapy Platform Advances Toward First-in-Human Study - Continuity Biosciences and Breakthrough T1D partner to develop an implantable, prevascularized chamber for islet cell transplantation with localized immune protection. – Continuity Biosciences
Immune Interventions at T1D Onset Show Promise - An NEJM editorial reviews modest but encouraging results from GAD-alum, vitamin D, and probiotic interventions at diagnosis, calling for larger trials. – Johnny Ludvigsson
Lived Experience of Youth with T1D - A qualitative review highlights how AID systems reduce burden and increase autonomy for young people, though access inequities persist across economic and educational lines. – PubMed
That's a wrap for this week. The algorithms are getting smarter, the data is getting richer, and the challenge now is making sure everyone who needs these tools can actually get them. Until next time, take care of yourselves.
Made with ❤️ by Data Drift Press - Hit reply with your questions, comments, or feedback. We'd love to hear from you!