Insulin Sensitivity Factor (ISF) Calculator
Insulin Sensitivity Factor (ISF) Calculator
The 1800 rule, the 1500 rule, and the two constants the published pump data actually support. None of these is a trial result: the 1500 rule was devised by one physician for regular insulin and the 1800 rule is a modification of it by textbook authors, while the only two retrospective studies that did the arithmetic properly arrived at 1700 and 1960. A starting point for a clinician to titrate, and nothing more.
Insulin sensitivity factor
Rule + total daily dose → ISFTotal daily dose 40 units/day, 1800 rule (rapid-acting analogue)
Formula
ISF (mmol/L per unit) = ISF (mg/dL per unit) ÷ 18
Correction dose (units) = (current glucose − target glucose) ÷ ISF
- the constant
- 1800 for a rapid-acting analogue and 1500 for regular human insulin as taught; 1700 (Davidson 2008) and 1960 (Walsh 2010) as fitted to real pump data. The spread between 1500 and 1960 is a 30% difference in the resulting correction dose
- total daily dose
- basal plus all bolus insulin over a typical 24 hours, averaged across at least a week. A single day’s total, or the prescribed rather than the delivered dose, moves the answer proportionally
- the mmol/L form
- 1800 ÷ 18.016 = 99.9 and 1500 ÷ 18.016 = 83.3, which is why “100 ÷ TDD” and “83 ÷ TDD” circulate. They are arithmetic restatements of the mg/dL rules and carry no independent authority
- what it is not
- not a measurement, not a trial endpoint, and not one number per person. It varies through the day within the same individual and between people on identical total daily doses
- insulin on board
- the correction formula above assumes no active insulin from an earlier dose. A rapid-acting analogue is still lowering glucose three to four hours after injection, and ignoring that is the commonest single cause of an iatrogenic low
Worked example
Total daily dose 40 units/day, 1800 rule (rapid-acting analogue)
1800 ÷ 40 = 45 mg/dL per unit
In mmol/L: 45 ÷ 18 = 2.5 mmol/L per unit
So a correction from 14 mmol/L down to a target of 7 mmol/L is (14 − 7) ÷ 2.5 = 2.8 units, call it 3
Now the same patient on the other three constants: 1500 ÷ 40 = 38 mg/dL (2.1 mmol/L), 1700 ÷ 40 = 43 (2.4), 1960 ÷ 40 = 49 (2.7)
That spread turns the same correction into anywhere from 2.6 to 3.3 units
Which is the point: the rule sets the order of magnitude, and the diabetes team sets the number, by watching what corrections actually do
The four constants, and where each comes from
| Constant | Insulin | Source | Trial-derived? |
|---|---|---|---|
| 1500 | Regular human insulin | Attributed to Paul Davidson, Atlanta — clinical practice, published through pump-education material | No |
| 1800 | Rapid-acting analogue | A modification of the 1500 rule by the authors of the pump-training books, made because glucose drops faster and further on the analogues. The same source notes some use 2000 | No |
| 1700 | Rapid-acting analogue (pump) | Davidson PC et al., Endocr Pract 2008 — least-squares fit to pump data from patients at target | Retrospective, peer-reviewed |
| 1960 | Rapid-acting analogue (pump) | Walsh J et al., J Diabetes Sci Technol 2010 — 1,020 pump downloads, 396 analysed | Retrospective, peer-reviewed |
ISF at a range of total daily doses (1800 rule)
| Total daily dose | ISF mg/dL per unit | ISF mmol/L per unit | Units to drop glucose by 5 mmol/L (90 mg/dL) |
|---|---|---|---|
| 20 units/day | 90 | 5.0 | 1.0 |
| 30 units/day | 60 | 3.3 | 1.5 |
| 40 units/day | 45 | 2.5 | 2.0 |
| 50 units/day | 36 | 2.0 | 2.5 |
| 60 units/day | 30 | 1.7 | 3.0 |
| 80 units/day | 23 | 1.2 | 4.0 |
| 100 units/day | 18 | 1.0 | 5.0 |
Sensitivity is not one number per day
| Time of day | Measured ISF, median (mg/dL per unit) | What follows |
|---|---|---|
| Morning | 50 (IQR 40–75) | The largest correction doses are needed here — the dawn rise in cortisol and growth hormone raises insulin requirement |
| Afternoon | 60 (IQR 40–75) | One unit does more than it did at breakfast |
| Evening | 75 (IQR 45–77.5) | The smallest correction doses. A morning factor applied in the evening over-corrects |
Where these constants came from, and how much to trust them
The insulin sensitivity factor, or correction factor, is how far one unit of rapid-acting insulin is expected to lower blood glucose. It is the number a pump’s bolus calculator or a person doing the arithmetic by hand uses to turn “my glucose is 14 and I want it at 7” into a dose. Getting it wrong in the direction of too small a factor means too much insulin, and too much insulin means hypoglycaemia. That asymmetry is why the provenance of the constant matters more here than the arithmetic does.
And the provenance is thinner than the confident round numbers suggest. The 1500 rule came first, for regular human insulin, attributed to Paul Davidson in Atlanta and disseminated through pump education rather than through a trial. The 1800 rule is a modification of it: the authors of the pump-training books raised the constant because glucose “tends to drop faster and farther” on lispro and aspart than on regular insulin, and the same source records that some practitioners use 2000 instead. Neither figure has a randomised trial, a validation cohort or a stated precision behind it. They are clinical judgement, written down.
Two retrospective studies did put numbers to it. Davidson and colleagues, publishing in Endocrine Practice in 2008, fitted the correction factor by least squares to pump data from patients at target and arrived at 1700 ÷ total daily dose, with basal insulin at 47% of the total. Walsh and colleagues, analysing 1,020 pump downloads in the Journal of Diabetes Science and Technology in 2010, arrived at 1960 ÷ total daily dose, with basal at 48%. So the empirical range is roughly 1700 to 1960, the teaching rule sits inside it, and the honest reading is that the rule gets the order of magnitude right and should not be trusted to two significant figures. A page that presented 1800 ÷ TDD as the formula, without any of this, would be presenting a teaching heuristic as a derivation — which is the error to avoid on a page whose output is an insulin dose.
The deeper problem is that sensitivity is not a single number per person at all. The one study to measure it against real correction doses through the day found it varying significantly within the same individual — lowest in the morning, when cortisol and growth hormone raise insulin requirement, and highest in the evening — and concluded that the 1800 rule fitted morning corrections and not afternoon or evening ones. The interquartile ranges in that study also overlapped heavily across time of day, meaning two people on similar doses differ from one another about as much as each differs across their own day. Illness, exercise in the previous 24 hours, menstrual phase, alcohol and glucocorticoids all move it too. So the rule gives a starting factor; the diabetes team then titrates it by watching what corrections actually do, usually with different factors for different parts of the day, and usually more than once. Pair this with the insulin-to-carbohydrate ratio calculator, which has the same provenance problem, and with the total daily insulin dose calculator, which supplies the denominator both of them depend on.
Frequently asked questions
What is the 1800 rule?
Insulin sensitivity factor in mg/dL per unit equals 1800 divided by the total daily insulin dose, for rapid-acting analogues. It is a modification of the earlier 1500 rule for regular human insulin, made by the authors of the pump-training literature rather than derived from a trial, and the same source notes that some practitioners use 2000 instead.
What is the mmol/L version of the 1800 rule?
Roughly 100 divided by the total daily dose. That is not a separate rule: 1800 divided by 18.016 is 99.9. The 1500 rule becomes 83 divided by the total daily dose by the same arithmetic. Both are restatements of the mg/dL rules in another unit.
Which constant should I use, 1500, 1700, 1800 or 1960?
1500 only for regular human insulin. For a rapid-acting analogue, 1800 is the usual teaching starting point, while the two retrospective pump studies support 1700 and 1960. The spread changes a typical correction dose by around 30%, which is why the number on the pump should end up being set from observed responses rather than from any of them.
Is the 1800 rule accurate?
It sets the right order of magnitude and no more. The only study to measure real-life sensitivity through the day found it appropriate for morning corrections but not for the afternoon or evening, and found sensitivity varying widely between people on similar total doses. Treat the output as a starting point for titration.
Why does my sensitivity factor change during the day?
Insulin requirement rises in the early morning with cortisol and growth hormone, so one unit does less at breakfast than at dinner. Measured medians in one paediatric cohort ran 50 mg/dL per unit in the morning, 60 in the afternoon and 75 in the evening. Many people end up with two or three different factors set for different parts of the day.
Can I change my correction factor myself?
This is a decision for you and the clinician who prescribes your insulin. Correction factors interact with the carbohydrate ratio, the basal dose and the insulin-on-board setting, so a change made in isolation can produce lows that look as though they came from somewhere else. Take the number to your diabetes team, change one setting at a time, and review it.
Related calculators
References
- Davidson PC, Hebblewhite HR, Steed RD, Bode BW. Analysis of guidelines for basal-bolus insulin dosing: basal insulin, correction factor, and carbohydrate-to-insulin ratio. Endocr Pract. 2008;14(9):1095–1101.
- Walsh J, Roberts R, Bailey T. Guidelines for insulin dosing in continuous subcutaneous insulin infusion using new formulas from a retrospective study of individuals with optimal glucose levels. J Diabetes Sci Technol. 2010;4(5):1174–1181.
- Hegab AA. Diurnal variation of real-life insulin sensitivity factor among children and adolescents with type 1 diabetes using ultra-long-acting basal insulin analogs. Front Pediatr. 2022;10:854972.
- American Diabetes Association Professional Practice Committee. 9. Pharmacologic Approaches to Glycemic Treatment: Standards of Care in Diabetes—2025. Diabetes Care. 2025;48(Suppl 1):S181–S206.
Medical Disclaimer: The tools and content provided here are for educational and reference purposes only. They are not intended to substitute for professional medical advice, diagnosis, or treatment. Clinical decisions should always be based on the comprehensive assessment of a qualified healthcare professional.