In this week’s issue of The Savvy Diabetic:
-
- Modular Medical wins FDA clearance for Pivot tubeless insulin patch pump
- I Asked AI to Count My Carbs 27,000 Times
- Popular AI chatbots are confidently dispensing medical misinformation
- WEBINAR: Diabetes Technology: Choosing What’s Best For You
- Sana Biotechnology/ Mayo Clinic Strategic Collaboration on T1D Care
- The Progress to the Cure with Carolyn Robertson,APRN, MSN, CDCES
- VIDEO: Conquering Fear and Embracing Life with Dr. Mark Heyman
- Early-stage funding slumps toward post-pandemic low
- Genetic risk for type 1 diabetes extends to brain cells
Modular Medical wins FDA clearance for Pivot tubeless insulin patch pump by Sean Whooley for DrugDeliveryBusiness.com, 9 April 2026.
Modular Medical announced that it received FDA 510(k) clearance for its next-generation insulin patch pump. Pivot, a tubeless patch pump, consists of two parts, with a 3 mL reservoir and a disposable battery. It features smartphone connectivity for bolus and monitoring, plus electronic accuracy without tubes or constant wear. The pump requires no battery recharging and has removability for showering, sports or discretion. Its quick bolus button enables insulin delivery without a separate controller.
San Diego-based Modular Medical targets “almost-pumpers” on multiple daily injections (MDI) with its insulin pump. It hopes to deliver an option for adults with type 1 or type 2 diabetes with hesitancy toward traditional pumps. The company believes Pivot offers simplicity, affordability and flexibility. It hopes to improve adoption and long-term outcomes while reducing costs.
By developing its patented insulin delivery technologies, the company hopes to improve access to glycemic control. Its founder, Paul DiPerna, previously founded leading insulin pump maker Tandem Diabetes Care. DiPerna invented and designed Tandem’s t:slim pump.
Read more: Modular Medical wins FDA clearance for Pivot tubeless insulin patch pump
I Asked AI to Count My Carbs 27,000 Times. It Couldn’t Give Me the Same Answer Twice. by Tim Street for Diabettech.com, 15 April 2026.
Ask ChatGPT to estimate the carbs in your lunch. Now ask it again. And again. Five hundred times. You’d expect the same answer each time. It’s the same photo, the same model, the same question. But you won’t get the same answer. Not even close — and the differences are large enough to cause a hypoglycaemic emergency. That’s the central finding of a study I’ve just published as a preprint, and it has direct implications for anyone using AI-powered carb counting in a diabetes app. Every model returned different carbohydrate estimates for the same photo across repeated queries. But the degree of disagreement varies enormously.
I found food identification errors in 8 of the 13 test images:
-
-
-
- Bakewell tart: Claude called it a “Linzer torte” in 100% of 510 queries. GPT-5.4 called it a “jam tart” or “cake bar.” Only Gemini 3.1 Pro correctly named it (99.8%).
- Crema catalana: Three of four models called it “creme brulee” 100% of the time. Only Gemini 3.1 Pro got “crema catalana” — in 3.4% of queries.
- Cheese sandwich: Gemini 3.1 Pro added non-existent “deli meat” in 17.4% of queries — hallucinating an ingredient that isn’t there. This could directly inflate carbohydrate estimates.
-
-
What this means for you: The DTN-UK stated earlier this year that generic LLMs must never be used as autonomous advisory calculators for insulin delivery. This data is the quantitative evidence base for that statement.
If you’re using AI carb counting in a diabetes app:
-
- Don’t trust it blindly. No model tested is safe for unsupervised insulin dosing. Not even Claude.
- Ask it more than once. Query 3-5 times and look at the spread. If the answers vary wildly, the model is uncertain — even if it doesn’t tell you that.
- Check what it thinks it sees. If the model identifies “chicken with stuffing” for your grilled fish, you want to know that BEFORE it calculates your carbs.
- Know your model. The four-fold difference in consistency is real. Claude Sonnet 4.6 is the safest single-query option today, but it can still be precisely wrong.
- An AI that gives you the same wrong answer 500 times is not more trustworthy than one that varies. Consistency is necessary but not sufficient.
Read more: I Asked AI to Count My Carbs 27,000 Times
Popular AI chatbots are confidently dispensing medical misinformation by Sadie Harley for MedicalXpress.com, 14 April 2026.
A substantial amount of medical information provided by five popular chatbots is inaccurate and incomplete, with half (50%) of the responses problematic: 30% were somewhat, and 20% were highly problematic. These are the results of a study published in the journal BMJ Open.
Continued deployment of these chatbots without public education and oversight risks amplifying misinformation, warn the researchers. Key findings on prompt type and models
-
-
-
- Prompt type was influential: open-ended prompts, for example, produced 40 highly problematic responses—significantly more than expected—and 51 non-problematic responses—significantly fewer than expected. The opposite was true of closed prompts.
- Reference quality was poor, with an average completeness score of 40%.
- Answers were consistently expressed with confidence and certainty, with few caveats or disclaimers.
-
-
Read more:
WEBINAR: Diabetes Technology: Choosing What’s Best For You by diaTribe.org, scheduled Monday, 27 April 2026.
Diabetes Technology:
Choosing What’s Best For You
-
-
-
- 27 April 2026
- 2:00-3:15pm EDT / 11:00am-12:15pm PDT
-
-
Continuous glucose monitors (CGMs). Automated insulin delivery (AID) systems. Smart pens. Inhaled insulin. Powerful data platforms. There are so many diabetes tech options out there today, and choosing the right one can make your life easier.
Our next Musings webinar will focus on how to choose diabetes tools that truly fit your daily life. Instead of a one-size-fits-all approach, our expert panelists will focus on what actually matters: your lifestyle, your preferences, and your goals. Whether you’re exploring new tools or rethinking your current setup, this discussion will provide practical guidance, honest perspectives, and greater confidence in choosing what works for you.
Our highly-regarded speakers include Dr. Viril N. Shah (Indiana University) and Amy Tenderich, MA (DiabetesMine). The panel will be moderated by Dr. Robert Gabbay (Harvard Medical School).
To Register: Diabetes Technology: Choosing What’s Best For You
Sana Biotechnology/ Mayo Clinic Strategic Collaboration Focused on Improving Care in T1D, announced by Sana
Sana Biotechnology, Inc., a company focused on changing the possible for patients through engineered cells, and Mayo Clinic today announced a strategic collaboration to advance development of SC451, Sana’s investigational hypoimmune-modified pancreatic islet cell therapy for type 1 diabetes. SC451 is designed to allow a single administration of pancreatic islet cells to support long-term glucose control without the need for ongoing insulin therapy or immunosuppression for patients with type 1 diabetes.
The collaboration will draw on Mayo Clinic’s multidisciplinary expertise to accelerate the development, validation, and standardization of protocols and processes for SC451, supporting safe, scalable, and consistent delivery across diverse clinical environments. These areas include:
-
-
-
- End-to-end clinical and operational insight to optimize workflows, including product handling and delivery of SC451, and post-treatment care to enable broader adoption across Mayo Clinic and global care settings.
- Surgical expertise, including refining procedural techniques.
- Standardizing handling, delivery and post-treatment management.
- Leadership in clinical trial design, including biomarker identification to guide patient selection and longitudinal monitoring.
-
-
Mayo Clinic will also look to advance its capabilities in the delivery of investigational islet cell therapies, further strengthening its leadership in innovative, multidisciplinary treatment approaches. In connection with the collaboration, Mayo Clinic will make an equity investment in Sana Biotechnology, reflecting a shared commitment to advancing innovative approaches aimed at improving care for patients with type 1 diabetes. The organization also has the option to make an additional equity investment under the terms of the agreement.
Read more: Sana Biotechnology and Mayo Clinic Announce Strategic Collaboration
The Progress to the Cure with Carolyn Robertson,APRN, MSN, CDCES for New York Diabetes Program.com, 17 April 2026.
Recently, there has been a shift in focus from the management of symptoms to the development of disease-modifying therapies and functional cures. The landscape is changing rapidly but the promised cure is not going to happen quickly. Though there have been multiple successes that suggest that it actually can happen.
Recent clinical results from 2025 and 2026 show breakthrough progress in preserving insulin production and achieving insulin independence through cell therapy. While these studies are in varying stages of development, there are several that have reached Stage 3 of the clinical trial pathway. These are likely to be submitted to the FDA for consideration in the not-too-distant future. Others are in the early stages of evaluation but the initial results are exciting. I have included this table to help explain the different phase of clinical testing.
Read more: Development of disease-modifying therapies and functional cures
Early-stage funding slumps toward post-pandemic low by Nick Paul Taylor for FierceBiotech.com, 16 April 2026.
A slow start to 2026 has put first-time biotech financings on course for their worst year since before the pandemic, reducing the already limited funding opportunities available for startups. The data come from J.P. Morgan’s first-quarter biopharma licensing and venture report (PDF). Analysts tracked 50 seed and series A investments collectively worth $2.3 billion over the first three months of 2026.
Read more: Early-stage funding slumps toward post-pandemic low
Genetic risk for type 1 diabetes extends to brain cells by Mahima Samraik, Yale University for MedicalXpress.com, 13 April 2026.
A study published in Nature Communications suggests that there may be a much closer link between these symptoms and the drivers of type 1 diabetes. Researchers found that the genetic variation associated with type 1 diabetes risk is active not only in immune and pancreatic cells, but also in certain brain cells.
“There are shared biological pathways between type 1 diabetes and brain function,” says senior author David A. Alagpulinsa, Ph.D., assistant professor of comparative medicine at Yale School of Medicine. “The function in the brain cells may help regulate immune thresholds.”
“Traditionally, researchers have studied the genetic basis of type 1 diabetes by focusing on immune cells and the pancreatic beta cells that those immune cells attack,” Alagpulinsa says. “We took a different approach. We asked whether the genetic signals that affect type 1 diabetes are also active in brain cells.””The most important finding in our work is the fact that the genetic risk for type 1 diabetes is also active in brain cells,” Alagpulinsa says. “This is one of the first studies tying microglia to type 1 diabetes genetic risk.”
Read more: Genetic risk for type 1 diabetes extends to brain cells
