How to Calculate a Meal's Insulin Load (Examples)
To calculate the insulin load of a meal, work out each ingredient’s score with 0.56 × net carbs + 0.1 × protein + 0.03 × fat (all in grams), then add the scores together. Net carbs means total carbohydrate minus fibre. The total is the meal’s Insulin Load Score: under 30 is low, 31 to 60 moderate, above 60 high on our scale. Below are the steps and three worked plates, with the maths shown.
By Naor Lugassi, founder of Kcaly AI. Last updated: November 6, 2026. Not medical advice, and not a tool for insulin dosing.
The formula, in one line
The Insulin Load Score (ILS) behind our insulin load calculator weights nutrients by how strongly they tend to drive insulin: carbohydrate most, protein a little, fat barely. It is a simplified estimate inspired by the Food Insulin Index, which was measured in people. If you want the reasoning first, read insulin load score and insulin load score explained.
ILS = 0.56 × (carbs − fibre) + 0.1 × protein + 0.03 × fat
Step by step
- List what is on the plate, with weights. Cooked weight for rice, pasta and oats; include the oil and sauce, which are easy to forget.
- Get carbs, fibre, protein and fat per ingredient. Use the pack label, or USDA FoodData Central for whole foods. Our insulin load pages already list these for about 70 common foods.
- Scale to the amount you ate. Values are usually per 100 g; 150 g means multiply by 1.5.
- Apply the formula to each ingredient. Subtract fibre from carbs first. Never let net carbs go below zero.
- Add the ingredient scores. Because the formula is linear, this equals applying it once to the plate’s total carbs, fibre, protein and fat.
- Compare against the bands. 0 to 30 low, 31 to 60 moderate, above 60 high. Use it to compare plates, not to diagnose anything.
A single ingredient by hand
Take 100 g of cooked white rice: 28 g carbs, 0.4 g fibre, 2.7 g protein, 0.3 g fat. Net carbs are 27.6 g. So 0.56 × 27.6 = 15.5, plus 0.1 × 2.7 = 0.3, plus 0.03 × 0.3 = 0.01, for about 15.7. Carbs supply almost all of it, which is why starch dominates the totals below.
Example 1: salmon, broccoli and olive oil
| Ingredient | Net carbs (g) | Protein (g) | Fat (g) | Score |
|---|---|---|---|---|
| Salmon, 150 g | 0.0 | 30.0 | 19.5 | 3.6 |
| Broccoli, 150 g | 6.6 | 4.2 | 0.6 | 4.1 |
| Olive oil, 10 g | 0.0 | 0.0 | 10.0 | 0.3 |
| Meal total | 451 kcal | 8.0 | ||
This plate lands at 8.0, well inside the low band. Salmon contributes through protein and a little fat; broccoli through a few net carbs. Olive oil adds calories but almost no score, a reminder that this number is not a calorie count.
Example 2: same salmon, swap in white rice
| Ingredient | Net carbs (g) | Protein (g) | Fat (g) | Score |
|---|---|---|---|---|
| Salmon, 150 g | 0.0 | 30.0 | 19.5 | 3.6 |
| White rice (cooked), 200 g | 55.2 | 5.4 | 0.6 | 31.5 |
| Broccoli, 100 g | 4.4 | 2.8 | 0.4 | 2.8 |
| Meal total | 606 kcal | 37.8 | ||
One swap and the total moves to 37.8: roughly 4.7 times the first plate, almost all from the rice. Calories also differ (606 vs 451 kcal), so a lower score is not automatically a better plate for every goal. Halving the rice would remove about half of that rice contribution.
Example 3: a “healthy” breakfast
| Ingredient | Net carbs (g) | Protein (g) | Fat (g) | Score |
|---|---|---|---|---|
| Oatmeal (cooked), 250 g | 25.8 | 6.3 | 3.8 | 15.2 |
| Banana, 1 medium | 23.9 | 1.3 | 0.4 | 13.5 |
| Almonds, 28 g | 2.5 | 6.0 | 14.0 | 2.4 |
| Meal total | 447 kcal | 31.1 | ||
Oats, banana and almonds look virtuous and total 31.1. Most of it is oats and banana. The almonds help (fibre is subtracted, fat barely counts), but they do not offset the starch and fruit sugar. If you want to lower it, shrink the oat portion or swap the banana for berries before cutting the nuts.
Comparing the three plates
| Meal | Calories | Insulin Load Score | Band |
|---|---|---|---|
| Salmon, broccoli, olive oil | 451 | 8.0 | Low |
| Salmon, white rice, broccoli | 606 | 37.8 | Moderate |
| Oatmeal, banana, almonds | 447 | 31.1 | Moderate |
For ideas on which ingredients keep the total down, see low insulin load foods list.
Where the arithmetic goes wrong
- Hidden fat and sugar. Dressings, cooking oil and sauces are the usual misses. Oil barely moves the score but sauces with sugar can.
- Raw versus cooked weights. Dry rice is about three times denser in carbs than cooked rice. Match the data to the state you weighed.
- Fibre double counting. Some labels already list net carbs. Subtract fibre only once.
- Protein-rich foods and dairy. The Food Insulin Index work (Holt, Miller and Petocz, 1997, PubMed) found some protein-rich foods raise insulin more than their carbs predict. A macro formula understates that, so treat low scores on these foods as a floor, not a guarantee.
Doing it without a spreadsheet
The maths is easy; the weighing, looking up and typing is what makes people quit. You can paste a plate’s totals into the calculator and get the score in one step. In our own use of Kcaly AI, the habit that sticks is sending the meal as a photo, text or voice note on WhatsApp and reading the score off the reply (an observation from using the product, not a study). The Insulin Load Score feature does the per-ingredient maths for you, and our calorie tracker for diabetes page covers the setup. Or start the 3-day free trial.
Sources
- Holt SHA, Miller JC, Petocz P. An insulin index of foods: the insulin demand generated by 1000-kJ portions of common foods. Am J Clin Nutr 1997;66:1264-76.
- Nutrient values: USDA FoodData Central (fdc.nal.usda.gov), via the Kcaly AI food database.
Frequently asked questions
How do you calculate the insulin load of a meal?
Calculate each ingredient's score as 0.56 × net carbs (carbs minus fibre) + 0.1 × protein + 0.03 × fat, using grams for the amount you actually ate, then add the ingredient scores together. The total is the meal's Insulin Load Score. It is an estimate, not a measured insulin response.
Do I add up the scores or calculate once for the whole plate?
Both give the same answer, because the formula is linear. You can add the per-food scores, or add up total net carbs, protein and fat for the plate and apply the formula once. Adding per food is easier when you want to see which ingredient drives the number.
What is a good insulin load for a meal?
On the Kcaly AI calculator, 0 to 30 counts as low, 31 to 60 as moderate and above 60 as high. These bands are our own convention for comparing plates, not clinical thresholds. What suits you depends on your goals, activity and any medical condition.
Is insulin load the same as carbs or glycemic load?
No. Glycemic load models the blood glucose effect of carbohydrate only. Insulin load also gives weight to protein and fat, which stimulate insulin to a smaller degree. For carb-heavy meals the two numbers move together; for high-protein, low-carb meals they diverge.
Can I use a meal's insulin load to dose insulin?
No. The score describes how much insulin demand a meal tends to create, on average, in a simplified model. It is not a dosing tool and must never be used to set or adjust insulin. Follow your diabetes care team's ratio and correction plan.
Share this with a friend
Send it straight to WhatsApp in one tap.
مستعد لتتبع أذكى؟
Kcaly AI يتتبع السعرات والبروتين والماكروز ومؤشر الحمل الأنسوليني — كل ذلك عبر واتساب.
ابدأ مجاناً — تجربة لمدة 3 أيام