CGM Guide Series
CGM Guide for Type 1 Diabetes
A nine-year-old going back to school with a new sensor and a parent following from work. A twenty-something on the 780G changing her sensor every Friday morning before the commute. A sixty-year-old whose adhesive question made her ask about the annual implantable. Three different fits, three different sensors. CGM in 2026 is no longer a search for the single most accurate device; the framework filters that question down to five mainstream sensors that all clear the bar. The choice that remains is fit-to-life: which AID system, which family-following workflow, which wear duration, which form factor. This guide walks you through the framework, the five mainstream devices, and the two we are watching but not yet recommending.
Ask Grace
Tell Grace what matters most to you. She walks you through the accuracy data, the AID compatibility, and the family-following workflow for each device, and helps you narrow to two before a clinic conversation.
The cluster at a glance
Five mainstream UK CGM systems, plus two on the watching list. The compact map below names what each row is and is not.
Five questions. Every CGM must answer all five.
This framework was built with the DSN Forum UK. It is not a checklist of features. It is a test of whether a CGM has been studied in a way that allows you to trust the reading when you are deciding how much insulin to take.
- People with type 1 diabetes included, studies in T1D, not just type 2 or healthy volunteers
- Meal challenge included, rapid glucose rises tested, not just stable periods
- Hypoglycaemia challenge included, accuracy tested in the low range where it matters most
- Venous or capillary blood reference, the gold-standard reference method, not another CGM
- Population and indication match, studied in the people it will actually be used in, with the right regulatory indication
The five mainstream devices
Five CGMs have published the evidence to clear the framework. Each has a dedicated page with the full accuracy picture, the trade-off it sits on, and the daily-use workflow that distinguishes it from the others in the cluster.
Device 1, Dexcom G7
5/5, ±20/20 95.3%, 10 days plus 12-hour grace. The widest AID compatibility in the cluster, paired with Omnipod 5, Tandem t:slim X2 with Control-IQ, and CamAPS FX. The G7 is what most of the cohort is now using when AID flexibility is the priority.
Device 2, FreeStyle Libre 2 and 3
5/5, ±20/20 93.4%, 14 to 15 days. Two AID lanes inside one family: Libre 2 Plus pairs with Omnipod 5; Libre 3 (and 3 Plus) pair with CamAPS FX on YpsoPump. The Libre 3 is the smallest CGM commercially available, and LibreLinkUp is the family-following app most often named first in paediatric clinic.
Device 3, Roche Accu-Chek SmartGuide
5/5, ±20/20 90.5%, hypo-range 94.3%, 14 days. The only CGM in this cluster carrying three labelled prediction layers (30-minute, two-hour, overnight). Adults 18 and over by labelling, with mandatory day-one calibration. No AID partnership at this time; the trade-off the device sits on is proactive overnight protection vs algorithm-driven insulin delivery.
Device 4, MiniMed Simplera Sync
5/5, ±20/20 88%, 6.5 days. The 780G-locked all-in-one disposable, factory calibrated, upper arm only. One of three sensors approved for the MiniMed 780G alongside Guardian Sensor 4 and the Abbott-made Instinct (15-day, CE Mark March 2026). Outside the 780G the Simplera Sync is not available; inside it, the choice is which 780G sensor fits the way the family lives.
Device 5, Senseonics Eversense 365
5/5, ±20/20 93.4%, outside ±40/40 0.4%, one implant per year. The only implantable CGM. Inserted under local anaesthetic into the upper arm, no patch on the skin, vibration alerts on the removable transmitter. Adults 18 and over, no AID integration. Page rebuild ahead in this cluster; framework profile complete.
Watching, not yet recommending
Two more devices are close to the framework threshold. Both have CE marking and manufacturer-led accuracy data; peer-reviewed publication and the wider evidence base needed to clear the framework are still pending. Until that lands, neither is recommended by GNL Grace and neither is included in the selector below.
What is in the watching list and why
CareSens Air (Spirit Health, distributed in the UK). CE marked; manufacturer-published accuracy data on file. Pivotal-trial publication and an independent head-to-head against the cluster’s reference comparators are the gating evidence step. Once peer-reviewed accuracy lands and the framework can be applied with confidence, the page will be rebuilt to flagship format and added to the mainstream row above.
GlucoMen iCan (Menarini Diagnostics). CE marked; published data is pending or under review. Same gating conditions as CareSens Air. The dedicated page will be rebuilt and added to the mainstream row when the evidence base is in.
This is not a judgement on either product; it is an honest position on what we can recommend right now from the published evidence.
Find your best match
Pick the three things that matter most to you. The selector scores each device against your priorities and recommends your top two.
Why only these devices? Every CGM in this selector has passed the five-question clinical framework on this page. Devices without published accuracy studies in the right populations, or without a non-adjunctive indication, are not included. The selector is a starting point for a conversation with your diabetes team, not a substitute for it.
Select 3 priorities (0 of 3 chosen)
What the accuracy numbers actually mean
Marketing material typically leads with a statistical average. The average tells you very little about what matters clinically. GNL reports CGM accuracy using ±20/20 agreement (the proportion of readings close enough to dose insulin from without a finger-prick) and ±40/40 agreement (the proportion within the outer safety band). Across the five mainstream devices, ±20/20 sits between roughly 88% and 95%; the share of readings outside ±40/40, the small fraction where errors become clinically large, varies between about 0.1% and 1.2%. The full thesis on how to read these numbers, and what to do about the small fraction outside the safety band, lives on the dedicated accuracy page.