Mealtime Insulin and Dosing Basics:
Why Timing Beats Dose
Everything in one place. Read the plain version with Jude, earn your way into the evidence with Grace, then the full model with John. Stop wherever you have enough.
How we teach: three rules, borrowed from Taleb
You earn each level by showing you understand it, not by scrolling past it. We only teach what we would use on ourselves and the people we love.
Understanding beats memory and luck, so the checks reshuffle every time you retry. A pass means you got it, not that you guessed it. And we teach you to tell a trend (signal) from one reading (noise).
We give you the scaffolding and get out of your way. Roam where your curiosity leads, go as deep as you want, and ask Grace anything. We will not teach a bird how to fly.
Want to understand your own meal patterns, in your units? Ask Grace, then take it to your care team.
One page, three depths
This guide compounds: each layer rests on the one beneath it. Read Jude’s plain version, then pass a short understanding check to open Grace, then another to open John. You can roam freely within a layer; you cannot skip ahead a layer, because the next one would not make sense and you would be standing on a gap.
The whole thing, in plain words
Mealtime insulin is the rapid-acting dose you give to cover the glucose that comes from food. It has one job: to be there in the bloodstream at the same time as the glucose from your meal. The hard part is not the amount; it is the timing. Food can raise glucose within minutes, but injected insulin takes its time to wake up, often peaking around an hour after the dose. So if you eat and dose at the same moment, the food is already racing ahead while the insulin is still putting its boots on.
That is why a small head start helps so much. Giving the dose around 15 to 20 minutes before most meals lets the insulin catch up with the food, and it tends to lower the after-meal spike a good deal. This is the single most useful habit in mealtime dosing, and it costs nothing.
The amount itself comes from a ratio: a personal number, agreed with your care team, that says how many grams of carbohydrate one unit of insulin covers for you. Fatty and high-protein meals are the tricky ones; a pizza or a takeaway can keep nudging glucose up for hours after a fast meal would have settled, so the same insulin given all at once may run out too soon. And once a dose is in, it keeps working for hours. That leftover, working dose is called insulin on board, and forgetting it is how people stack one dose on another and go low. Mealtime insulin is not personal advice; your ratios and your timing rules are set with your diabetes care team.
Does this match the life of the person living it? Nobody pre-boluses perfectly at every meal, and a restaurant that is slow to bring the food will catch you out. The aim is a habit you can keep on a normal day, not a rule that makes you anxious every time you eat.The Pemberton lens, lived recognisability, one of the four GNL appraisal lenses.
The numbers underneath
Why food beats insulin off the line
Rapid analogue insulin (Humalog, NovoRapid, Apidra) typically begins working at around 15 minutes, peaks at 60 to 90 minutes, and keeps lowering glucose for around 4 to 5 hours.1 Glucose from a fast meal arrives much sooner than that. The gap between the two curves is the structural reason post-meal spikes happen even when the dose is exactly right.2 Two things drive it: insulin from a jab or pump absorbs slowly from under the skin, and it never reaches the liver in the concentration the body’s own insulin would, so the liver’s mealtime glucose output is less well damped.
A schematic of the timing mismatch, not measured patient data. Curves are illustrative shapes; the point is the early gap when food has arrived and insulin has not yet caught up.
What a pre-bolus actually buys
Closing that early gap is what pre-bolusing does. In a paediatric pump crossover trial (Cobry 2010), giving the same calculated dose 20 minutes before the meal produced significantly lower glucose at one and two hours than dosing at the meal or after it; the after-meal arm gave the highest peak of all.3 A narrative review pulling the pharmacology together found mealtime insulin taken 15 to 20 minutes before eating cut post-meal glucose by roughly 30 percent with less hypoglycaemia than dosing at the bite (Slattery 2018).2 Paediatric guidance carries the 15-to-20-minute pre-bolus for most meals at its strongest evidence grade.4 Faster insulins (Fiasp, Lyumjev) need a shorter lead or none; very sugary meals can need longer.
| Insulin family | Example | Peak | Action duration | Typical pre-meal lead |
|---|---|---|---|---|
| Ultra-rapid | Fiasp, Lyumjev | earlier than rapid | ~3 to 4 h | short or none |
| Rapid | Humalog, NovoRapid, Apidra | 60 to 90 min | ~4 to 5 h | 15 to 20 min |
| Regular (older) | Actrapid, Humulin R | 2 to 3 h | ~6 to 8 h | longer, 30 to 45 min |
Pharmacokinetics from Leohr 2021 (URLi versus Humalog clamp, N=190) and Heinemann 2002. Do not generalise lead times across analogues; the family you use changes the answer.1
The meals that break the rule: fat and protein
Carbohydrate is not the whole story. Fat and protein add their own, later glucose rise, from around 2 hours out to 5 hours. A systematic review found high-fat, high-protein meals consistently needed more insulin and a different delivery shape (Bell 2015).5 In one closed-loop study, a high-fat dinner needed 42 percent more insulin than a low-fat one at identical carbohydrate (Wolpert 2013), and a controlled trial found high-fat, high-protein meals needed on average around 65 percent more insulin (range 17 to 124 percent) delivered as a split over roughly two and a half hours (Bell 2016).5 The mechanism is delayed gastric emptying: fat slows the carbohydrate into the late window, where a standard single bolus has already faded. The population-level response is to shorten the pre-bolus and spread the dose (a split or extended bolus) so more insulin lands late. These are population averages; the size of the uplift and the split for any one person and meal is a care-team conversation.
“65 percent more insulin” sounds precise until you read the range beside it: 17 to 124 percent. The honest headline is the direction, not the decimal. When a single number is offered for something as variable as a meal, ask how wide the spread around it really is.The Goldacre lens, evidence-grade discipline, one of the four GNL appraisal lenses.
Insulin on board, and the patience it demands
Insulin on board is a model, not a measurement
Once a dose is given it keeps working for hours, and the amount still active is called insulin on board (IOB). The IOB figure on a pump or automated system is a calibrated model output, not a reading of the insulin in your body.6 The model rests on an assumed action duration; on most automated systems (MiniMed 780G, Tandem Control-IQ, Omnipod 5, CamAPS FX) you cannot edit that duration directly, and it is the single strongest lever on how the algorithm behaves. The practical consequence: device IOB can read near zero while a meaningful fraction of a meal dose is still physiologically active, which matters most around exercise.7 Understanding this is more leveraged than any insulin switch.
The stacking trap, and why patience is the safer move
A rapid-analogue correction takes around 3 to 4 hours to land most of its effect, and glucose often does not move much in the first 30 to 60 minutes.8 That delay is where the most common mealtime safety failure lives: seeing a still-high number, assuming the first dose did not work, and adding a second one inside the action window of the first. The two then compound, and the low arrives hours later. The population-level lessons are: account for what is still on board before adding more, and remember that a 10-to-20-minute walk is often a quicker lever than a second dose when the CGM trend allows.8
The hardest part of correcting is waiting long enough to judge whether the first dose worked. Most of a correction’s effect is still to come at the one-hour mark. Any rule about how long to wait, and any figure for how much to give, belongs with your diabetes care team, not a web page.
Correction figures, framed the only honest way
It is tempting to want a single number: a glucose level in, a dose out. GNL does not give that, on principle. Every individual has their own personal correction factor, set with their care team from their total daily dose, their history, their age, and their hormonal and activity patterns. A figure built from a population-average rule (such as a fixed number divided by total daily dose) cannot stand in for it. So where correction is discussed anywhere across GNL, the figure is framed as a population-average estimate at a given total daily dose, never a personalised dose, and you are reminded to apply your own correction factor.9 If you do not know your correction factor, the figure is not for you to act on; that is a conversation with your team. This is a locked clinical-safety position, not a stylistic choice.
The asymmetry is everything. An under-correction you can fix at the next reading; an over-correction stacked on insulin still on board can take you somewhere you cannot afford to go. When the downside is unbounded and the upside is small, the patient move is the safe one. Protect hardest against the low you cannot undo.The Taleb lens, robustness to outliers, one of the four GNL appraisal lenses.
An IOB display is a model standing on an assumption you usually cannot see or change. Treat it as a useful instrument with a known blind spot, not as ground truth, and name the blind spot out loud. A figure is only as trustworthy as the assumptions underneath it; never sell the model as the territory.The Hayes lens, technical and methodological rigour, one of the four GNL appraisal lenses.
The whole guide, summarised
Glucose never lies; it just records what the meal and the insulin did, in their own time. Learn the timing, respect what is still working, and set the numbers with your team.
This page is the taster. The full journey on Mealtime Insulin, three modules and their 30 questions, with your progress saved, lives in Learn with Grace. Glucose never lies; come and learn to read it.
References
Evidence grades A (strongest) to D (editorial or working analysis).
- Leohr J, et al. Pharmacokinetics and glucodynamics of ultra rapid lispro (URLi) versus Humalog: a pooled analysis. Clin Pharmacokinet. 2021;60(8):1015-1027 (N=190 clamp); analogue absorption per Heinemann L. Variability of insulin absorption and insulin action. Diabetes Technol Ther. 2002;4(5):673-682. A
- Slattery D, Amiel SA, Choudhary P. Optimal prandial timing of bolus insulin in diabetes management: a review. Diabet Med. 2018;35(3):306-316. A (GNL anchor on pre-bolus principle; narrative review.)
- Cobry E, et al. Timing of meal insulin boluses to achieve optimal postprandial glycemic control in patients with type 1 diabetes. Diabetes Technol Ther. 2010;12(3):173-177 (paediatric pump crossover, n=20). A
- de Bock M, Codner E, et al. ISPAD 2024 Clinical Practice Consensus Guidelines: Nutritional Management (Chapter 10), pre-bolus 15 to 20 minutes at Grade A. Horm Res Paediatr. 2024. A
- Bell KJ, et al. Impact of fat, protein, and glycemic index on postprandial glucose control in type 1 diabetes: systematic review. Diabetes Care. 2015;38(6):1008-1015; Wolpert HA, et al. Dietary fat acutely increases glucose concentrations and insulin requirements. Diabetes Care. 2013;36(4):810-816; Bell KJ, et al. Optimized mealtime insulin dosing for fat and protein. Diabetes Care. 2016;39(9):1631-1634. A
- GNL synthesis on insulin on board as a model output, anchored to device documentation and the IOB pharmacokinetics evidence pack. D GNL clinical synthesis on a Grade A/B base.
- Bassi M, et al. evidence on closing the IOB configuration gap (TIR gain without hypo increase); GNL IOB trade-off synthesis. D GNL synthesis on a Grade A/B base. [VERIFY-MA]
- Correction-insulin pharmacodynamics (3 to 4 h action, delayed onset) and stacking risk, GNL clinical synthesis anchored to Leohr 2024 hyperglycaemia recovery RCT and the GNL Compliance Dossier Section A5. D GNL synthesis on a Grade A/C base.
- Correction Dose Framing Policy (HARD RULE, locked 1 May 2026, John Pemberton), gnl-grace/wiki/policies/correction-dose-framing.md. D GNL clinical-safety policy.
One page, three voices: Jude, Grace, John. Population-average, not personalised.
Keep learning: Diabetes, the foundations · HbA1c and Time in Range · CGM guide · Hypoglycaemia · Exercise · Automated insulin delivery (AID)