Automated insulin delivery, CGM performance data, the growing Type 2 diabetes evidence base, and current HSE referral and funding pathways — summarised for referring GPs and endocrinology colleagues.
Automated insulin delivery (AID) is now standard of care for Type 1 diabetes and has an accelerating evidence base in insulin-treated Type 2 diabetes. Current-generation CGM accuracy (MARD ~8%) supports insulin dosing decisions in most circumstances. This review, authored by Dr Syed Kazmi, Consultant Endocrinologist and Diabetologist at Naas Cardiology & Endocrinology Clinic (IMC 213626), summarises the current device landscape, key 2025–2026 trial data, and the practical referral/funding pathway for patients in Ireland.
Three hybrid closed-loop systems dominate current UK/Ireland and European practice: Medtronic MiniMed 780G, Tandem t:slim X2 (Control-IQ+), and Insulet Omnipod 5. All use predictive/model-based algorithms to auto-adjust basal insulin and deliver automatic correction boluses; none yet fully automate meal-time dosing.
| System | CGM pairing | Time-in-range change | Notes |
|---|---|---|---|
| Medtronic MiniMed 780G | Guardian sensors | +21.1% | 780G is a closed ecosystem; the newer sibling MiniMed Flex (FDA-cleared March 2026, screenless, half the size) breaks this pattern — pairs with Simplera Sync and Abbott's Instinct sensor, and is also cleared for insulin-requiring Type 2 diabetes |
| Insulet Omnipod 5 | Dexcom G6/G7, Libre 2 Plus, Libre 3 Plus (added 2025–2026) | +15.2% | Tubeless; broadest paediatric use; first system cleared for Type 2 diabetes (August 2024) |
| Tandem t:slim X2 (Control-IQ+) | Dexcom G6/G7; Libre 3 Plus (from July 2026) | +10.1% | Touchscreen; sleep/exercise activity modes. Libre 3 Plus support live in UK, Switzerland, Sweden, Finland, Italy — not yet confirmed for Ireland at time of writing |
| Parameter | Dexcom G7 | FreeStyle Libre 3 / 3 Plus |
|---|---|---|
| Adult MARD | ~8.2% arm-placed (pivotal trial) | ~7.9% (Libre 3 pivotal trial; no separate Plus-specific figure identified) |
| Warm-up | 30 min | 60 min |
| Sample rate | Every 5 min (288/day) | Every 1 min (1,440/day) |
| AID integration | Broadest — Tandem Control-IQ+, Omnipod 5 | Expanding — Omnipod 5 (2025–26); Tandem t:slim X2 from Jul 2026 (UK/EU markets; not yet confirmed for Ireland) |
| HSE LTI list | List B (Dexcom G6/G7) | List A (Libre 2); Libre 3 Plus not yet on List A |
Both sensors remain within accuracy bands considered suitable for insulin dosing without routine confirmatory fingerstick testing, consistent with manufacturer labelling. Device choice for a given patient is more often determined by pump compatibility and HSE reimbursement list status than by the marginal MARD difference. Manufacturer comparison data.
Until recently, AID was used almost exclusively in Type 1 diabetes. That has changed rapidly over 2025–2026:
Clinical implication: insulin-treated Type 2 diabetes patients with recurrent hypoglycaemia, high glycaemic variability, or HbA1c persistently above target on optimised MDI now represent a reasonable population for specialist AID assessment, pending local funding availability. Note the TIR benefit in the 2026 meta-analysis is graded low-certainty evidence — reasonable to discuss with patients as an emerging rather than fully established benefit. Systematic review and meta-analysis — PMC.
For patients on multiple daily injections who are not pump candidates or do not wish to use one, connected pens (NovoPen 6/Echo Plus, InPen) provide automated dose-timing data and bolus calculation, improving adherence visibility at review. A Swedish real-world study found NovoPen 6 use associated with up to two additional hours per day in target glycaemic range compared with standard pens — useful data to discuss when a patient is ambivalent about pump therapy but has adherence or dose-timing concerns. Diabetes Ireland.
Patients increasingly ask about smartwatch/smart-ring "glucose monitoring." Clinicians should be aware the FDA issued an explicit safety communication on 21 February 2024 stating no such device is approved, cleared, or authorised for glucose measurement, and that inaccurate readings can lead to incorrect insulin or sulfonylurea dosing with potentially severe consequences (confusion, coma, death). This is a useful point to proactively raise, particularly with patients considering discontinuing an approved CGM in favour of a wearable claim. Genuine non-invasive sensing (Biolinq Shine — FDA De Novo classification 23 September 2025 — uses a shallow intradermal microsensor array, not a needle/filament) remains positioned for non-insulin-dependent Type 2 diabetes use at present, not as an insulin-dosing-grade replacement for CGM. FDA Safety Communication.
GP referrals can be sent via Healthmail or the standard referral pathway — see referral information for details.
Dr Kazmi accepts GP and specialist referrals for CGM, insulin pump, and AID assessment, including insulin-treated Type 2 diabetes. Referrals via Healthmail preferred.