1. ISO 15197 floor

The 15/15 regulatory entry ticket. Every meter on this page clears it.

ISO 15197:2013 requires 95% of readings within ±15% of reference at glucose at or above 5.55 mmol/L, and within ±0.83 mmol/L at glucose below. Every meter in this canon passes that floor. That makes it the entry ticket to the comparison, not the differentiator within it.

2. The ±5% band

The proportion of readings within ±5% of reference. Spread across the 18 meters is 46.6% to 86.7%. This is where the clinical decision sits.

±5% is the band that decides whether a correction-dose decision at a low reading is safe or unsafe. The 40-percentage-point spread across NHSE Cat 2 meters means the choice of meter is a genuine clinical decision, not a procurement coin-toss. Surfacing the 15/15 band alone obscures the spread that matters most clinically; the 5/5 band is the one to read.

3. Strip longevity and oxidation resistance

Maximum strip life from manufacture, open-pot expiry rule, enzyme chemistry contribution.

Strip life is not glamorous; it is the difference between a fresh reading at three months and a pot you cannot trust at six. Open-pot rules across this canon range from no-discard-date (use to the printed pot expiry) to four months after opening, which compresses the useful life of a 50-strip pot considerably for light users. Glucose dehydrogenase (GDH) chemistry is less oxygen-sensitive than glucose oxidase (GOx); for supplemental-oxygen users, arterial sampling, and users at altitude this matters, though empirical accuracy in the headline table can override the chemistry assumption.

4. Haematocrit tolerance range

The wider the published range, the more robust the meter at the extremes.

Anaemia, dehydration, and polycythaemia all sit outside a typical adult haematocrit window; a meter with a narrow tolerance range loses accuracy at exactly the populations where readings carry the most weight. The published range across this canon runs from 0 to 70% at the widest to 35 to 50% at the narrowest. The lower bound is the practical flag: a 35% floor excludes anaemic patients from the meter's validated accuracy envelope.

5. Long shelf life

Manufacture-to-expiry interval. Procurement and light-use cohorts care; high-frequency users do not.

Shelf life is the interval from strip manufacture to the printed pot expiry. It matters for primary-care procurement (stockpiles that sit before being dispensed) and for the light-use cohort where the pot can expire before the strips are used. The longest published maximum strip life in this canon is 24 months; most pack inserts do not publish the figure at all, and the open-pot rule (criterion 3) is usually the more material constraint.

6. Second Chance Sampling

Add more blood to the same strip without re-lancing. Reduces failed tests in older-adult and paediatric cohorts.

Second Chance Sampling lets the user add more blood to the same strip within seconds when the initial drop is insufficient, removing the need to lance again. The clinical impact lands hardest in older adults with thin skin, in children with small fingers, and in any user where each failed test is a wasted strip and a second lance. Only one meter in the NHSE Cat 2 canon currently carries this feature; the rest require a fresh strip and a fresh lance on every failed sample.

+ Ketone capability (helpful add-on)

Same handheld accepts both blood glucose strips and blood ketone strips. Material for T1D sick-day and pump-failure protocols.

Dual-purpose meters take one handheld where two would otherwise be required. For T1D users this lands during illness (sick-day protocols), during pump failure (DKA-risk monitoring), and during any clinical scenario where carrying one device beats two. The subset of NHSE Cat 2 meters that accept both strip types is small (four meters; one not yet launched); the trade-off worth naming is that dual-purpose convenience can come with accuracy below the canon median, and the right call on that trade-off depends on how often the user actually tests ketones.

The standout case

Where dual-purpose blood-ketone testing is not part of the user's routine, the framework points straight to the top of the ranking table. The decision rule is: read the ±5% column; pick the meter at or near the top of the table that matches local formulary access. Reasoning: the band that decides correction-dose safety at a low reading is the ±5% band, and the spread on that band across the 18 commissioned meters is wide enough that meter choice is a genuine clinical decision, not a procurement coin-toss. The eat-your-own-dogfood test holds here: this is the meter we would use ourselves, give to our family, give to the people we care about.

The dual-purpose case

Where blood-ketone testing on the same handheld is part of the user's routine (T1D sick-day, pump failure, DKA-risk), the ketone-subset table narrows the shortlist. The decision rule then is: read the ±5% column within that ketone subset; pick the meter that meets the user's haematocrit range and connectivity requirements. The trade-off worth naming is real: dual-device convenience comes, in the lowest-accuracy ketone-capable meter, at the cost of accuracy below the canon median, and that is a clinical conversation, not a default. The framework asks how often the user actually tests ketones (occasional sick-day use versus frequent monitoring under high-risk protocols); the answer to that question is what tips the trade-off one way or the other, and that conversation belongs with the user's care team.