Automated insulin delivery (AID) systems have come a long way, and with more options than ever, clinicians can no longer treat all pump algorithms as interchangeable. Nurse practitioner Erin Henry, presenting at the Association of Diabetes Care & Education Specialists annual meeting, emphasized that understanding the specific algorithm powering a patient's pump is foundational to effective diabetes management — from adjusting carb-to-insulin ratios to troubleshooting site failures.

One of the most critical distinctions Henry highlights is between total daily dose (TDD) algorithms (like Omnipod 5, where the pump sets the basal rate and manual changes have no effect) and adaptive basal algorithms (like Tandem or twiist, where clinician-set basal rates actively influence outcomes). Choosing the wrong approach — or not knowing which type a patient uses — can mean interventions that do nothing at all.

Key Takeaways:

• For type 1 diabetes, make smaller (10%) insulin rate changes due to higher insulin sensitivity; for type 2, 20% changes are appropriate
• When starting a GLP-1 alongside insulin, expect ~30% drop in insulin requirements for T1D and ~20% for T2D — adjust proactively
• Review AID data every 3 months for stable patients, every 4–6 weeks for those with poor control, and weekly during pregnancy
• Focus on time-in-automation, total daily dose, and basal-to-bolus ratio as the most actionable metrics

Why it matters: With AID systems becoming standard of care, clinicians who don't understand algorithm-specific behavior risk giving patients advice that's ineffective — or worse, harmful. Knowing the system means knowing how to truly help.