Part 3 explored the main ways diabetes devices calculate IOB, which systems use which approach, the strengths and limitations of each model, and how people often work around those limitations in practice.
Part 4 looks ahead. It explores two future models that could help people with Type 1 diabetes get the best of both worlds: a display that more accurately reflects true physiological insulin exposure, alongside the flexibility to choose how aggressive, balanced, or protective correction behaviour should be.
That matters because current IOB is being asked to do two different jobs at the same time.
- Represent physiological insulin exposure
- Regulate correction behaviour
This is the central design conflict running through current IOB systems.
When one number is expected to do both, trade-offs become unavoidable.
So the real question is not whether IOB matters — it clearly does.
The question is whether one number can safely control correction behaviour and represent physiological insulin exposure at the same time.
Option 1 is pragmatic and evolutionary.
Option 2 is structurally cleaner, but more ambitious.
Option 1: Keep the correction model and add physiological exposure as a separate entity
This approach keeps the existing bolus calculator architecture intact.
Active Insulin Time (AIT/DIA) would simply be reframed for what it already behaves like in practice:
a correction aggressiveness dial.
- Shorter AIT/DIA (2–3 hours)
- Less IOB deducted sooner
- Corrections allowed earlier
- More aggressive behaviour
- Longer AIT/DIA (4–6 hours)
- More IOB deducted for longer
- Corrections restricted
- More conservative behaviour
This preserves what current systems already do well:
- The current evidence base
- Regulatory familiarity
- Clinical understanding
- Existing device behaviour
Alongside this behavioural dial, a second model could run in parallel in the background.
Physiological insulin exposure (U/kg-based)
This model estimates how much insulin effect may still be circulating in the body.
Instead of using a fixed decay assumption, exposure is estimated using:
- Recent insulin doses
- Body weight
- Units per kilogram (U/kg)
- Dose-dependent insulin action profiles (as discussed in Part 1)
The aim is not perfect prediction.
Crucially, this type of IOB model;
- would not block corrections (+)
- would not alter the bolus calculator (+)
- would inform of activty hypo risk
Why?
Insulin on Board and Exercise Hypoglycaemia Risk
Large real-world datasets from T1DEXI consistently show that insulin on board at the start of exercise is one of the strongest predictors of hypoglycaemia risk.
| IOB at start of exercise | Events (%) | Hypoglycaemia risk |
|---|---|---|
| <1 U | 3,874 (44%) | 6% |
| 1–<2 U | 1,788 (20%) | 9% |
| 2–<3 U | 1,191 (13%) | 10% |
| ≥3 U | 1,974 (22%) | 12% |
Even with current modelling limitations, a clear dose–response relationship is visible: as insulin exposure rises, hypoglycaemia risk increases.
However, the insulin exposure in these analyses was calculated using a 4-hour linear decay model based on absolute units, not units per kilogram.
This creates two major problems.
- Body weight is ignored, meaning the same number of units represents very different insulin exposure between individuals.
- Linear decay underestimates physiological insulin action, particularly:
- the rising effect during the first 60–90 minutes, and
- the longer tail of larger boluses.
As a result, the true relationship between insulin exposure and exercise hypoglycaemia risk is almost certainly underestimated.
Importantly, even with these limitations the signal is already visible in the data.
This suggests that if insulin exposure were modelled using:
- units per kilogram
- time-stamped boluses
- physiologically realistic insulin action curves
- longer exposure windows (e.g. up to 8 hours)
the dose–response relationship between circulating insulin and hypoglycaemia risk would likely be substantially stronger.
As someone involved in two analyses examining exercise glycaemia using the T1DEXI datasets (Glucose Go and Glucose Low), John Pemberton of The Glucose Never Lies, notes;
“We improved the IOB model by units per kilogram were used. However, the simplified 4-hour linear decay model we used still massively underestimated true insulin exposure.“
Key implication
The existing datasets are extremely valuable, but the way insulin exposure is currently represented is likely diluting its true impact.
Re-analysing these data using a physiologically based insulin exposure model, such as the one outlined in this guide and implemented in the Exercise IOB Calculator in Part 5, could reveal a much clearer understanding of how insulin drives exercise hypoglycaemia risk.
Why a physiological IOB model could be implemented safely
A key advantage of the physiological insulin exposure model proposed in this guide is that it can inform risk without interfering with insulin delivery systems. If implemented within diabetes technology platforms or research tools, its role would be informational rather than prescriptive.
- It would not block correction doses
- The model would not prevent or restrict user-initiated correction boluses.
- Existing insulin delivery systems would continue to function exactly as they currently do.
- Users and clinicians would retain full control over insulin dosing decisions.
- The model simply estimates circulating insulin exposure, allowing users to understand the potential risk environment for exercise or activity.
- It would not alter the bolus calculator
- The bolus calculator within pumps or apps would remain unchanged.
- Carb ratios, correction factors, targets, and algorithm logic would continue to operate exactly as they do today.
- The proposed model sits alongside existing dosing tools, rather than replacing them.
- Its purpose is not to calculate insulin doses, but to provide a physiologically meaningful estimate of insulin exposure.
- It would inform risk
- The model provides a clearer estimate of circulating insulin relative to body weight (U/kg).
- It accounts for time-stamped boluses and realistic insulin action curves, improving the representation of insulin exposure.
- This information can then be used to estimate hypoglycaemia risk during exercise or activity.
- Users and clinicians can make more informed decisions about carbohydrate intake, activity timing, or insulin adjustments.
Taken together, this means the model functions as a decision-support layer.
- It does not control insulin delivery.
- It does not change dosing calculations.
- It simply improves the understanding of insulin exposure and risk.
For research and device development, this creates a strong opportunity. A shared, physiologically grounded IOB model could allow existing datasets to be reanalysed consistently and could help both clinicians and people with type 1 diabetes better understand the relationship between insulin exposure and exercise hypoglycaemia risk.
To illustrate this concept we built a simplified exposure model that we incoprotated into the GNL Exercise IOB Calculator for T1D.
The model estimates circulating insulin intensity based on doses taken in roughly the previous eight hours, body weight, and dose-dependent insulin kinetics.
Exposure is then displayed using a graded scale:
- 0.00 U/kg – Very Low
- 0.02 U/kg – Low
- 0.04 U/kg – Med/Low
- 0.06 U/kg – Med
- 0.08 U/kg – Med/High
- 0.10 U/kg – High
- 0.12 U/kg – Very High
- 0.14 U/Kg -Extremley High
This model does not predict glucose levels.
Instead it provides a physiologically consistent estimate of circulating insulin exposure.
In the spirit of The Glucose Never Lies®, the purpose is learning rather than perfect prediction.
Different exposure levels interact with different activities, training states and environmental conditions.
Over time that pattern recognition becomes extremely powerful.
If you want to explore this concept in practice, you can use the Glucose Never Lies® Exercise IOB Calculator for T1D.
Option 2: Redesign the architecture — separate meal and correction insulinand model IOB physiologically
This approach is structurally cleaner.
Instead of pooling all insulin into a single IOB bucket, separate insulin according to intent.
Step 1: Model expected glucose behaviour at beals
At meal entry the system would estimate the expected glucose rise based on factors such as:
- Carbohydrate type
- Meal composition
- Pre-bolus timing
- Post-meal activity
This creates a simple glucose behaviour profile rather than a precise prediction.
Meal insulin would then be handled within that model rather than automatically restricting corrections.
Step 2: Track Correction Insulin Separately
Correction insulin would be tracked independently using the same U/kg-weighted exposure model described above.
- Dose intensity (U/kg)
- Dose-dependent duration
- Curvilinear decay profiles
This governs stacking and correction behaviour without meal insulin artificially limiting future corrections.
Step 3: Display total physiological exposure
otal insulin exposure — including both meal and correction insulin — should ideally be modelled internally and then displayed in a clear, intuitive way that is easy for anyone to understand. Presenting exposure simply as a units/kg value is informative, but it is mainly useful for clinicians or those deeply engaged in diabetes physiology.
For most users, a graded risk scale would be far more practical. A simple RAG-style exposure indicator (low → moderate → high insulin exposure) allows people to quickly understand their current insulin load and the potential implications for activity or exercise.
This approach is illustrated in the GNL Exercise Insulin on Board Calculator for Type 1 Diabetes, where insulin exposure is translated into an easy-to-interpret visual scale. By combining a clear exposure signal with practical guidance, users can see the full picture of their insulin load rather than relying on a single abstract number.
This is particularly useful for situations where insulin exposure matters most:
- Exercise planning
- Overnight risk
- Repeated correction cycles
The key distinction becomes simple:
- Correction insulin influences correction behaviour
- Total physiological exposure informs hypoglycaemia risk
This removes the structural conflict present in current systems.
Which path is more likely?
Option 1 is more likely in the near term because it fits existing device architecture.
It fits existing device architecture and could be implemented without rewriting correction logic.
Option 2 is more ambitious.
It is biologically cleaner but would require would require substantial redesign, validation, and regulatory work.
However, both approaches would represent an improvement over the current situation.
Understanding how insulin exposure behaves is the first step toward designing better systems.
If you want to explore this idea in practice, you can try the Glucose Never Lies® Exercise IOB Calculator for T1D.
Conclusion
This guide has highlighted three key realities about insulin on board.
First, physiological insulin exposure lasts longer than most models assume. When expressed as units per kilogram, rapid-acting insulin can continue to influence glucose for four to eight hours, particularly after larger boluses.
Second, current device models are trying to do two different jobs with one setting: enabling effective corrections while also protecting against hypoglycaemia. A single IOB model cannot optimise both perfectly, which is why every system involves trade-offs between aggressive, balanced, or more protective behaviour.
Third, we now have the ability to model insulin exposure far more realistically. Using time-stamped boluses, body weight, and physiological action curves allows insulin exposure to be represented in a way that better reflects real biology.
The model presented in this guide,, and implemented in the the Glucose Never Lies® Exercise IOB Calculator for T1D, is one example of how this can be done. It may not be perfect, but it represents a substantial improvement over simplified four-hour linear decay models.
Importantly, this type of modelling does not need to alter existing dosing systems. It can simply be used to inform risk, particularly during exercise, where circulating insulin exposure is often the dominant driver of hypoglycaemia.
Looking ahead, there is a clear opportunity to revisit large datasets such as T1DEXI using a more physiological insulin exposure model. Doing so may reveal a much clearer relationship between insulin exposure and exercise hypoglycaemia risk.
Standardising how insulin exposure is represented would benefit research, device development, and clinical practice alike.
Acknowledgements
A special thank you to Simon Helleputte, MSc, PhD (Faculty of Medicine and Health Sciences, Ghent University) and Joseph Henske, MD, FACE (Director of the Diabetes Program, Professor of Medicine, Division of Endocrinology and Metabolism) for their generous feedback and thoughtful input during the development of this guide.
I am also deeply grateful to Professor Michael Riddell (York University, Canada) for his ongoing support, patience with my very lengthy emails, and his ability to continually steer the discussion back to the most important question: “So what?” His perspective has been instrumental in helping refine the practical implications of insulin-on-board modelling and why it matters for people living with type 1 diabetes.s.
I am also grateful to colleagues who proofread parts of the guide and offered helpful feedback.
Finally, a huge thank you to the GNL Team, whose support, creativity, and friendship continue to make this work possible.
- Anjanee Kohli — Creative Director and my rock
- Professor Othmar Moser — my brother from another mother
- Professor Dessi Zaharieva — my sister from another mister
- Professor Adrian Brown — long-time dancing partner in the world of metabolic research and evidence-based practice
Thank you for reading.
The IOB Guide for T1D
- Hub: The Insulin On Board Guide for T1D
- Part 1 – The Insulin On Board–Physiology Mismatch
- Part 2 – Different Models For Calculating Insulin On Board
- Part 3 – Choosing a Device-Specific Insulin On Board Settings: What Are You Optimising For?
- Part 4 – The Future of calculating Insulin On Board: combining correction behaviour and exercise hypoglycaemia risk
- Part 5 – GNL Exercise Insulin on Board Calculator for T1D
- Part 6 – Reccomended Reading and Resources
