What endocrinologists are seeing as automated insulin delivery becomes routine: stronger time in range, high algorithm trust, and persistent clinical work around meals, exercise, data review, patient education, and access.
Audience: Endocrinologists and Diabetes Specialists
Countries: USA, UK, Canada, France, Italy, Germany
Survey records: 8773636
-Hero findings
– Quick Read — Key Findings
The full report follows the gap between reliable automation and the everyday exceptions endocrinologists still have to manage.
The 2026 ADA Standards describe automated insulin delivery as a preferred insulin delivery method for many people with insulin-requiring diabetes, while emphasizing individualized device selection, education, troubleshooting, and continued access.
That framing fits the MDForLives survey. The specialists here are largely familiar with AID and are no longer debating the category itself. They are judging what happens after adoption: whether the patient stays engaged, whether the data are usable, and how often the algorithm still needs human help.
Context: American Diabetes Association, Standards of Care in Diabetes – 2026, Diabetes Technology.
62.2% point to increased time in range as the most consistent benefit they observe with closed-loop use, far ahead of any other single outcome.
What the pattern means: The strongest perceived benefit is the one clinicians can see across the day. TIR is functioning as a practical summary of how well therapy handles everyday glucose exposure.
45.9% report significantly better engagement and confidence, but 29.7% still see only moderate improvement with ongoing education needs.
What the pattern means: Automation does not remove self-management. It changes its shape. The strongest responses still leave room for education, troubleshooting, and patient-specific readiness.
48.6% say closed-loop data are useful but still require interpretation. Only 37.8% describe the workflow as seamless and clearly actionable.
What the pattern means: AID data have crossed the threshold from availability to usefulness, but interpretation remains part of the clinical work for the largest group.
64.9% call current algorithms generally reliable with occasional manual intervention. Only 27.0% express high confidence across most clinical scenarios.
What the pattern means: The dominant position is neither distrust nor full autonomy. It is supervised trust: the system usually works, but the clinician expects exceptions.
32.4% identify meals and exercise as the biggest limitation, making everyday glucose dynamics the most common real-world weak point.
What the pattern means: The leading limitation is physiological and behavioral timing, not lack of algorithmic sophistication in the abstract.
75.7% choose fully automated meal detection and dosing as the single advancement that would most improve clinical value.
What the pattern means: The priority is not more automation as a label. It is removing the meal-related task that still defines hybrid closed-loop use.
For endocrinologists, the next phase of automated insulin delivery is less about proving that the technology can improve glucose outcomes and more about reducing the moments that still demand rescue, interpretation, or repeated education. The survey points to a familiar clinical reality: a system can be highly automated and still create meaningful work when meals, exercise, sensor lag, missed boluses, or data review fall outside the smooth path.
Use TIR and device summaries to identify the specific situations where automation repeatedly loses control, especially meals, exercise, sensor lag, and late or missed boluses.
Teach patients what to do when automation is interrupted, when meals are missed, how activity modes work, and when a pattern deserves clinician review.
A technically strong system can still underperform in practice if its data take too long to interpret or its training burden is too high.
Direct answers to the questions healthcare professionals are most likely to ask about these findings.
