Guide series, Part 3 of 5
Is Your Time in Range What You Think It Is?
The international consensus says aim for 70% time in range. But that target was designed around one type of CGM accuracy. Some devices systematically read higher time in range than others for the same physiological glucose. Until you know which calibration zone your CGM sits in, your TIR number is harder to interpret against the international 70% target.
Ask Grace
Want to know how accurate your sensor is in the range you spend most time in? Ask Grace.
Why this matters before anything else in the guide
Time in range (TIR) is reported as a single number from your CGM. It looks objective. It is not. Your TIR depends on the calibration behaviour of the device measuring you, and different consumer CGMs do not all calibrate the same way.
Two people with identical underlying glucose patterns can record TIR numbers that differ by 5-10% or more, simply because they wear different sensors. If you do not know how your device behaves, comparing your TIR with the international 70% target is less reliable. That is the gap this Part exists to close.
The headline: the international consensus 70% TIR target (Battelino et al., Diabetes Care 2019;42(8):1593-1603, doi 10.2337/dci19-0028, Grade D; carried forward by ADA 2024) was modelled on Zone P CGM behaviour. If your device sits in Zone B, reading below the physiological corridor, your TIR reading is systematically higher than the consensus assumed. The same physiological glucose pattern that reads 70% TIR on a FreeStyle Libre 3 would read roughly 75% on a Simplera.
The Zone A, P, B framework
Declared before you read it. John Pemberton, who writes this page and founded The Glucose Never Lies, is an author on Pemberton et al 2026, the clinical opinion this page teaches. So this is an explanation of his own framework rather than an independent assessment of it. His published competing interests read: “JP reports consultancy fees and speaker honoraria from Abbott, Dexcom, Medtronic, Insulet, and Roche. Member of the IFCC Working Group for CGM. Founder of The Glucose Never Lies Limited.” The paper is published and you can read it and disagree.
The A-P-B model comes from Pemberton et al 2026, an international clinical opinion published in Diabetes, Obesity and Metabolism. It describes the direction in which a CGM’s glucose estimates sit relative to blood glucose. It is not a quality ladder: A, P and B are not grades, and a device does not climb from one to the next by getting better.
The reference is not a single line. Capillary glucose rises earlier and peaks higher after a meal; venous glucose sits lower and lags behind. The gap between the two averages around 10% across a day, narrowing to roughly 5% at rest and widening to roughly 30% during rapid post-meal transitions, and it can reverse during a fast fall from a high. Those two traces bound a physiological corridor, with venous glucose generally forming the lower edge and capillary glucose the upper edge. True blood glucose sits somewhere inside it.
Picture a river rather than a line on a map. Capillary glucose is one bank, venous glucose the other, and the true value is somewhere in the water between them. The banks are close together when you are sitting still and a long way apart in the half hour after a meal. Asking “is this sensor right?” is asking the wrong question. The question is which side of the river it is standing on.
A CGM senses interstitial glucose and reports an estimate of blood glucose. Where that estimate lands depends on the reference dataset the manufacturer trained the algorithm on (capillary, venous or arterialised-venous), how in-vitro sensor sensitivity was mapped to in-vivo behaviour, and the adaptive filtering applied on the body. So the question the model asks is not “how good is this sensor?” but “where in the corridor, or outside it, does this system put its estimates?” Three answers.
Zone A, Above
Glucose estimates sit consistently above the corridor, which means above capillary blood glucose. This is a real, occupiable position, not a ceiling and not an aspiration. A system reading in Zone A shows you a higher number than your blood glucose, which flatters nothing: it makes hyperglycaemia look worse and hypoglycaemia look rarer than it is.
Zone P, Physiological corridor
Glucose estimates sit inside the corridor, between capillary and venous blood glucose. This is the alignment that matches typical peripheral blood-glucose behaviour, and it is the range the older CGM outcome trials were built on, which is why it is the reference point for the consensus targets rather than the “best” zone.
Zone B, Below
Glucose estimates sit consistently below the corridor, which means below venous blood glucose. A system reading in Zone B under-reports hyperglycaemia, so more readings fall inside the time-in-range window and the TIR number on the screen goes up for the same underlying glucose.
What each zone means in the clinic
Pemberton et al 2026 sets out the clinical reading of each zone directly. The interpretation below follows that table.[1]
| Zone | Clinical interpretation | Average bias of CGM values relative to reference glucose |
|---|---|---|
| A | Estimated glucose consistently above physiological blood glucose may prompt earlier or more frequent correction doses and may reduce the apparent exposure to hypoglycaemia. | Above capillary blood glucose |
| P | Estimated glucose within the physiological corridor reflects alignment with typical peripheral blood-glucose behaviour and corresponds to the range underpinning legacy outcome evidence for CGM devices. | Between capillary and venous blood glucose |
| B | Estimated glucose consistently below physiological blood glucose may generate more conservative insulin-dosing signals and may under-report hyperglycaemia. | Below venous blood glucose |
Which devices have actually been placed
Three systems, and only three. The head-to-head evidence behind the zone placements comes from a small set of manufacturer-independent studies in which adults with type 1 diabetes wore several sensors at the same time, on the same body, so the comparison is within the same person rather than across trials: Eichenlaub 2025 (Journal of Diabetes Science and Technology), its companion analysis Freckmann 2025 (Diabetes Care), and Sanfilippo 2025 (Diabetes, Obesity and Metabolism, during structured exercise). All three studies evaluated the same three systems: Abbott FreeStyle Libre 3, Dexcom G7, and Medtronic Simplera.
- All three systems fell within Zone P for time below range.
- Two of them, FreeStyle Libre 3 and Dexcom G7, stayed within Zone P across most of time in range and time above range.
- One, Medtronic Simplera, predominantly occupied Zone B for time in range and time above range, reading lower glucose levels on average.
- None of the three occupied Zone A. That is a finding about these three sensors, not a statement that Zone A is unreachable.
Every other CGM on the market is not yet not yet placed. If your sensor is not one of the three above, you cannot be sure what Zone it reads in. This is the main reason why we really need CGMtesting to be standardised
What a device needs to qualify for the zone adjustment, and which are not yet placed
Why most CGMs aren’t placed in the CGM-zone adjustment yet. Zone placement in Part 4’s CGM-zone section requires publicly available head-to-head accuracy data covering dynamic glucose conditions (rapid post-meal rises, exercise drops), measured against a stated reference method. Being calibrated to capillary glucose is a clue about where a system is likely to sit; it is not a placement.
- Dexcom G6 and FreeStyle Libre 2 come from the same manufacturers as two placed sensors, but neither was in the head-to-head studies. Sharing a maker is not evidence of sharing a zone.
- MiniMed Guardian 4 is a different sensor from Simplera and has not been evaluated head-to-head alongside it. Simplera’s Zone B placement does not transfer to it. (MiniMed is the spun-off company carrying the former Medtronic Diabetes business.)
- Roche SmartGuide is calibrated to capillary glucose, which is consistent with sitting in the corridor, but does not yet have comparable head-to-head dynamic data published.
- Dexcom Stelo is calibrated to a plasma reference, and its use case (type 2 diabetes and prediabetes wellness, not T1D risk prediction) puts it outside the scope of this adjustment in any case.
- Eversense (implantable) is excluded pending sufficient comparative data.
The full device-by-device comparison sits in the CGM Guide. The IFCC working group (Pleus 2025) has published a recommended framework for evaluating CGM performance, including how to test sensors under fast glucose change. It is a set of recommended testing procedures, so that the risk is understood, not a certification scheme, and there is nothing yet for a manufacturer to be certified against. The GNL position is that it should become a requirement for regulatory clearance internationally. See Part 4 for that argument in full.
What the bias actually does to your number
A Zone B device reads below the corridor, so it under-reports hyperglycaemia: values that were genuinely above 10.0 mmol/L are reported inside the range instead. The result is that the proportion of values falling between 3.9 and 10.0 mmol/L (the standard time-in-range window) goes up, even though nothing has changed physiologically. The number on the screen is higher; the underlying glucose is identical.
It is the speedometer that reads five miles an hour slow. You hold the needle under seventy the whole way, feel entirely law-abiding, and arrive having driven at seventy-five. The driving did not change. The dial did.
The gain comes almost entirely from the top of the range, not the bottom. In the head-to-head studies, all three evaluated systems sat within Zone P for time below range, so they broadly agreed about hypoglycaemia; they diverged on time in range and time above range. A device that reads low does not make your lows disappear, and reading a Zone B device as though it were “kinder” across the board is the wrong reading of it.
The approximate effect, for matched users with similar glucose patterns[2]:
| Underlying physiological TIR | Zone P device reading | Zone B device reading |
|---|---|---|
| ~60% | ~60% | ~65% |
| ~70% | ~70% | ~75% |
| ~80% | ~80% | ~85% |
Why this matters clinically
Comparing yourself to consensus targets
The 70% TIR target was set against the average behaviour of Zone P CGMs in trial cohorts. If you wear a Zone B device and you hit 70% TIR, your underlying physiological TIR is closer to 65%. You are further from the consensus than your number suggests. Conversely, hitting 75% on a Zone B device is roughly equivalent to 70% on Zone P.
Comparing yourself to a friend or family member
If two people compare TIR numbers and one wears a FreeStyle Libre 3 while the other wears a Simplera, those numbers are not on the same scale. The Simplera number is around 5 percentage points higher for matched glucose. This is not a flaw in either device; it is a calibration design choice. But it matters when people use TIR to compare or compete.
Switching devices
If you move from a Zone P device to a Zone B device, your TIR will likely jump by around 5 percentage points overnight, with no actual change in your glucose pattern. The reverse is also true. Knowing this prevents you (and your care team) from over-interpreting a device-driven shift as a clinical change.
Setting your personalised target
Part 4 of this guide gives you a personalised TIR target in two separate steps: your glycator status first, the stronger-evidenced calculation, then a CGM-zone adjustment on top, a weaker-evidenced, expert-consensus layer. The CGM zone is the easier of the two to establish, when your device has been placed at all: if you wear one of the three sensors named above, that sets your zone adjustment; if you do not, the glycator step stands on its own.
What the underlying research shows
Two papers behind the zone framework, and what each one found
Pemberton JS, Wilmot EG, Barnard-Kelly K, Leelarathna L, Oliver N, Randell T, Taplin CE, Choudhary P, Adolfsson P. CGM accuracy: contrasting CE marking with the governmental controls of the USA (FDA) and Australia (TGA), a narrative review. Diabetes, Obesity and Metabolism 2023;25 (doi 10.1111/dom.14962, PMID 36585365) identified that CE Marking is not a quality standard for CGM. Published online December 2022, in print 2023; it is cited in both forms in the literature.
Pemberton et al 2026 (Diabetes, Obesity and Metabolism 2026;28(4):2551-2565), an international clinical opinion with 21 authors, set out the A-P-B model and named the three things that decide where a system sits: the reference dataset it was trained on, how sensor sensitivity was mapped from bench to body, and the filtering applied on the body. It also explains why the differences are hardest to ignore under dynamic conditions: the corridor itself widens from around 5% at rest to around 30% during rapid post-meal transitions, so there is more room for two systems to disagree exactly when the reading matters most. The differences are not confined to those moments, though. Eichenlaub 2025 and Freckmann 2025 found consistent bias directions and materially different glycaemic metrics across ordinary two-week wear, not only during in-clinic challenges.
Until manufacturers disclose calibration architecture publicly, the A-P-B zone model (Pemberton 2026) is the working tool for interpreting any metric (including TIR) that is calculated from CGM data. The Via Negativa principle applies: remove the assumption that all CGMs read the same before adding any interpretation to the number on the screen.
Find your CGM’s zone
The full device-by-device classification, including UK availability, AID compatibility, and accuracy data, sits in the CGM Guide.
Notes
Notes [1] and [2], collected from the sections above
- Clinical interpretive model for calibration alignment, after Pemberton et al 2026, Diabetes, Obesity and Metabolism 2026;28(4):2551-2565, Table 2. Reproduced for education; the wording is the authors’. ← back to text
- These are approximate, framework-level differences. Individual variation exists. The 5-percentage-point figure is the central estimate; head-to-head accuracy data (CGM Pack 1) places the upper bound of the Zone B / Zone P gap nearer 10 percentage points for some matched pairs, so the real-world spread between two users on different devices may be larger than the zone table suggests. The point is the direction and the order of magnitude, not a precise number. ← back to text
Part 3 of 5
Is Your Time in Range What You Think It Is?