Two numbers that define a person's insulin dosing
Intensive insulin therapy replaces a fixed sliding scale with two personal constants. The insulin sensitivity factor (ISF, also called the correction factor) is how far one unit of rapid-acting insulin lowers blood glucose. The insulin-to-carbohydrate ratio (ICR) is how many grams of carbohydrate one unit covers. Every bolus is then just two divisions added together: how far above target you are divided by the ISF, plus how many grams you are eating divided by the ICR.
Both constants scale inversely with how much insulin a person uses in total. Someone on 100 units a day is roughly twice as insulin-resistant as someone on 50, so each unit does about half as much work. That is the entire logic behind the rules of 1800 and 500: divide a fixed constant by the total daily dose and you get a starting estimate that is right for that person's resistance.
The correction component is what a nurse computes at the bedside when a pre-meal reading comes back high, and it is a standard NCLEX and PA-school calculation. The meal component is what a person with type 1 diabetes does several times a day. Both live inside every insulin pump's bolus wizard, which is why the arithmetic is worth understanding rather than trusting blindly — when a pump gives an unexpected answer, it is usually because one of these settings is stale.
Where 1800, 1500 and 500 come from
These are empirical constants, fitted to observed insulin action rather than derived from physiology. Paul Davidson's work in the 1980s produced the 1500 rule for regular human insulin: divide 1500 by the total daily dose and you get the mg/dL fall per unit. When rapid-acting analogues arrived, their sharper, shorter action profile produced a larger apparent drop per unit and the constant was revised upward to 1800. Some clinicians use 2000 for people who are especially sensitive.
ISF = 1800 ÷ TDD. At a TDD of 50 units, ISF = 36 mg/dL per unit. At 100 units it halves to 18. The inverse relationship is the whole point: the more insulin someone needs in a day, the less each individual unit accomplishes.
ICR = 500 ÷ TDD. At a TDD of 50 units, ICR = 10 g per unit — the familiar 1:10 ratio. At 25 units it becomes 1:20; at 100 units, 1:5. The constant 500 is loosely tied to the roughly 500 g of carbohydrate a day that a typical mixed diet supplies, but treat it as an empirical starting point, not a derivation.
The correction. (BG − target) ÷ ISF. If the glucose is 240 and the target is 120, you are 120 mg/dL above target, and at 36 mg/dL per unit that needs 120 ÷ 36 = 3.33 units. Note that this expression is signed: below target it returns a negative number, which correctly reduces the meal bolus rather than being ignored.
The meal bolus. carbs ÷ ICR. Sixty grams at 1 unit per 10 g is 6 units.
The basal split. A widely used starting assumption is that about half the total daily dose is basal insulin and half is bolus. That is a titration starting point only; real splits range from about 40% to 60% basal, and a person whose basal is badly set will show it as glucose drift between meals rather than as a bolus problem.
In mmol/L countries the same rules are stated with different constants because the units differ by a factor of 18.0182: 1800 ÷ 18.0182 ≈ 100, so ISF in mmol/L per unit is 100 ÷ TDD, and the 1500 rule becomes 83 ÷ TDD.
Worked example: 50 units a day, glucose 240, eating 60 g
An adult with type 1 diabetes uses about 50 units of insulin aspart a day. The pre-lunch fingerstick reads 240 mg/dL, the prescribed target is 120 mg/dL, and lunch contains 60 g of carbohydrate. The pen delivers half units.
- Insulin sensitivity factor. 1800 ÷ 50 = 36 mg/dL per unit. One unit is expected to drop the glucose by 36 mg/dL.
- Insulin-to-carb ratio. 500 ÷ 50 = 10 g per unit, a 1:10 ratio.
- How far above target. 240 − 120 = 120 mg/dL.
- Correction component. 120 ÷ 36 = 3.33 units.
- Meal component. 60 ÷ 10 = 6.00 units.
- Total before rounding. 3.33 + 6.00 = 9.33 units.
- Rounded to the pen. 9.33 ÷ 0.5 = 18.67, which rounds to 19 half-units, so 9.5 units.
- Expected result. The 3.33 correction units should bring 240 down toward 120, and the 6 meal units should hold the post-meal rise, provided the carbohydrate count is right and the dose is given far enough ahead of the food.
Change one thing and watch the sensitivity. If the same person's total daily dose were 100 units rather than 50, the ISF would halve to 18 and the ICR to 5, so the same reading and the same meal would need 120 ÷ 18 + 60 ÷ 5 = 6.67 + 12 = 18.67 units — exactly double. The formula scales linearly with TDD, which is why a stale TDD produces a proportionally wrong dose in both components at once.
How to judge whether the answer is reasonable
Check the correction against the total daily dose. A single correction bolus larger than about a fifth of the whole day's insulin is a warning sign, not an instruction. It usually means the TDD used in the calculation is out of date, the basal rate is wrong, or something acute is going on — an infection, a failed infusion site, a missed dose.
Insulin on board is the biggest thing this arithmetic ignores. Rapid-acting insulin keeps working for four to five hours. A correction given two hours after the last one will stack with insulin still active and can drive the glucose low several hours later. Pump bolus wizards subtract active insulin before recommending a dose; a hand calculation does not, which is why the usual teaching is to leave three to four hours between corrections.
A negative correction is a real result. Below target the expression returns a negative number, and the clinically correct action is usually to reduce the meal bolus by that amount rather than to give the full meal dose. This calculator reports the negative correction and holds the total at zero if it would go below zero, because no device can deliver negative insulin.
Verify the factors against observed results, not the formula. The rules give a starting point. The test of an ISF is what actually happens: give a correction with no food on board, check four hours later, and if the glucose lands well below target the ISF is too small (too much insulin per mg/dL) and should be raised. The test of an ICR is a post-meal check two to three hours after eating with no correction involved.
Targets are individual. ADA guidance sets glycaemic targets by person and circumstance rather than by one number; a common pre-meal target for adults is 80–130 mg/dL, but a higher target is deliberately chosen for people with hypoglycaemia unawareness, limited life expectancy, or a history of severe lows.
Starting ISF and carb ratio by total daily dose
| TDD (units/day) | ISF, rapid-acting (mg/dL per unit) | ISF, regular (mg/dL per unit) | Carb ratio (g per unit) | Basal at 50% (units/day) |
|---|---|---|---|---|
| 20 | 90.0 | 75.0 | 25.0 | 10 |
| 30 | 60.0 | 50.0 | 16.7 | 15 |
| 40 | 45.0 | 37.5 | 12.5 | 20 |
| 50 | 36.0 | 30.0 | 10.0 | 25 |
| 60 | 30.0 | 25.0 | 8.3 | 30 |
| 75 | 24.0 | 20.0 | 6.7 | 37.5 |
| 100 | 18.0 | 15.0 | 5.0 | 50 |
| 120 | 15.0 | 12.5 | 4.2 | 60 |
Because 1800 > 1500, the rapid-acting rule always returns the larger sensitivity factor at the same TDD, which means a smaller correction dose for the same glucose excursion.
This is a study and cross-check tool, not a prescription
Insulin is one of the drugs most often implicated in serious medication errors, and a dosing decision depends on far more than the four numbers on this page: insulin already on board, renal and hepatic function, illness, steroids, exercise in the last several hours, pregnancy, gastroparesis, and the person's own history of hypoglycaemia. Use this calculator to learn the arithmetic and to check a figure you have already derived. Follow the prescriber's written orders and your institution's protocol for the actual dose, and confirm any unexpected result with a second clinician before giving it.
Where the arithmetic goes wrong in practice
- Using a stale total daily dose. Every output scales inversely with TDD, so a TDD that is 30% out makes both the correction and the meal bolus 30% out in the same direction.
- Applying the 1800 rule to regular insulin. Regular insulin uses 1500, which gives a smaller ISF and therefore a larger correction for the same excursion. Using 1800 by mistake under-doses.
- Mixing mg/dL and mmol/L. A glucose of 13.3 mmol/L entered as 13.3 mg/dL looks like a severe low. Use the unit selector rather than converting by hand.
- Double-correcting. Stacking a second correction on top of insulin still active from the first is the most common route to iatrogenic hypoglycaemia in people who dose by hand.
- Counting carbohydrate badly. The meal component is usually the larger of the two, so an error in grams matters more than the same proportional error in the glucose reading. Fibre and sugar alcohols are handled differently by different protocols.
- Dosing after the meal instead of before. Rapid-acting insulin still needs roughly 15 minutes of lead time to match the glucose rise from food; the same units given after eating give a higher peak and a later low.
- Treating the 50% basal split as fixed. It is a titration starting point. Persistent overnight or between-meal drift points at the basal, and no bolus formula will fix it.
Where this sits among dosing methods
The older alternative is the sliding scale: a fixed table of units by glucose range, identical for everyone on the ward. It is simple and it is reactive — it treats a high reading after it has happened and takes no account of what the person is about to eat or how insulin-sensitive they are. Basal-bolus dosing with an individual ISF and ICR replaced it in most inpatient and outpatient guidance precisely because the two constants personalise the response.
For inpatient work, weight-based starting doses are common: a TDD estimated as roughly 0.4–0.5 units per kg per day for an insulin-naive adult, then split half basal and half bolus, and titrated from there. That estimate feeds straight into the rules on this page. The same weight-based reasoning underlies the weight-based dosage calculation used for most other drugs.
Insulin is also an infusion drug. DKA and perioperative protocols run intravenous regular insulin at a rate in units per hour, computed and adjusted from hourly glucose measurements rather than from ISF — the same territory as an IV infusion rate and its drip-rate conversion. On the pharmacy side, the same insulin is dispensed in units per vial or pen, which is what a days supply calculation has to work from, and the electrolyte replacement that accompanies DKA treatment is quantified with milliequivalents.
Key terms
- Total daily dose (TDD)
- Every unit of insulin taken in 24 hours, basal plus bolus. It is the single input both rules divide into.
- Insulin sensitivity factor (ISF)
- Also called the correction factor. The expected fall in blood glucose from one unit of rapid-acting insulin, in mg/dL or mmol/L per unit.
- Insulin-to-carbohydrate ratio (ICR)
- Grams of carbohydrate covered by one unit. Written as 1:10, meaning one unit per 10 g.
- Insulin on board (IOB)
- Insulin from earlier boluses still acting. Rapid-acting analogues act for roughly four to five hours, and ignoring IOB is the main cause of stacked doses.
- Basal insulin
- Background insulin covering hepatic glucose output between meals and overnight, delivered as a long-acting injection or a pump's basal rate.
- Correction bolus
- The insulin given to bring an out-of-range glucose back to target, separate from any coverage for food.