How To: Do Dr. Morrison’s Carb Weighting System

This section is for everyone who wants to “cheat” on their low carb diet and minimise the consequences.  If you can see yourself happily low carbing without the need to cover higher carb meals you may skip to the How To: Time Insulin Injections for Simple Insulin Regimes.

You need to wean yourself gradually off your high carb diet.

You need to be in a typical transition period at the very least.

You need to adopt the 7 unit per shot guidelines.  Every single shot.

Your need to find out which methods of carb counting work best for you and do them every meal.

You need to keep your basals, exercise, and fat and protein intake consistent over the testing period.

You need to be well and free of infection or undue stress so your meal profiles will be accurate.

You need to test at 3am on an experiment day to see you have not over done your insulin.


What you are going to do is find out at what level your insulin stops working in a linear way.

You will recognise this by high blood sugars before the next meal compared to when you eat low carb meals. I don’t know when this will kick in for you. For Steven it was good bs at 30g and higher bs at 40g for the same carb insulin ratio.

The best meal to test on is your lunch. Your dawn phenomenon is not active. Your dusk phenomenon is not active. Your carb sensitivity is usually at its best. You are awake and can deal with any adverse effects on your blood sugars by correction doses before your evening meal.

It is too risky to experiement on yourself at your evening meal. Throughout this experimentation process the only sugars that were consistently perfect for Steven were the before bed ones for this reason.

You need to add a smaller incremental dose of insulin to the one calculated for your carb sensitivity for that meal.

Test yourself on items that are easy to calculate eg breakfast cereals that are easily weighed and bread slices that are listed on the package. Packaged processed food with carb labels are helpful for these experiments.

Decide on how many units of insulin extra you will give per 10g over the baseline figure.

Make this a very low amount. If you can change to a half unit pen. The novonordisk demi pens and junior pens have this facility and humalog has just become available in half units too.  Of course this is not a problem with a syringe.  For a child you may wish to consider extra accuracy from diluted insulin.

We started on 0.5 units novorapid for each 10g extra increase in carb.

If your figures are showing that this is giving sugars higher than your next pre-meal target you can up the amount of additonal carb weighting insulin by a small amount.

If your figures are showing that the next pre-meal bs is too low then up the amount of carb stages eg in 5 or 10g increments till you figure out what works.

This carb weighting method is accurate for Steven up to 90g of carb per meal. After this our levels are inaccurate.  They are usually lower than expected on this exponential weighting but sometimes are higher.

You could find that you need to start carb weighting at levels of less than 40g.

You could find that you don’t need to start carb weighting till levels of 50g or higher.

You may find half unit increments  need to be used at additional carb levels of  5g, 10g or 15g or 20g levels.

You will need to determine when this system stops being accurate for you. This could be at levels considerably lower than 90g or could perhaps be higher.

Remember to only do the experiments when conditions are optimal for this.

You are well.
Your carb levels are accurate.
You are starting at normal bs.
It is lunch time.
No unusual exercise is involved.
You are not particularly stressed.
You have help to figure out what you are doing.

By a slow process of guess and test you can find out how to extend the carb in your meals and still get normal bs levels before the next meal.

I would again stress that low carbing is the safest option.

Once you know how to deal with higher carbs at lunch safely you can test this out at other meals.

Once you have learned this method use it wisely. It is for emergency situations and special indulgences. If you use it day in and day out you may indeed have normal bs levels before your meals and at bedtime, but you will be spiking a lot more than any non diabetic will.  It is not only high blood sugars but widely fluctuating blood sugars that are causes of complications.

Quick Quiz:
There is no quiz for this section.

Where to Next?
Please continue on to the How To: Time Insulin Injections for Simple Insulin Regimes section.

Reference Info:

NOTE: the calculator is no longer available, but the methodology below may be of interest for those interested in creating their own app or tool (e.g. spreadsheet).

The insulin calculator in the download section of this site is a great help AFTER you have done your personal experiments and calculations.

It is a simple programme with three background tables that can be adjusted by you if your insulin sensitivities change.

The first factor which you adjust for are your insulin sensitivity at breakfast, lunch, dinner and bedtime. This has been covered on a previous module on the course.

The second factor which you adjust for is the carb weighting factor. This module shows you how to do this.

The third factor which you adjust for is your correction bolus for certain high blood sugars. For simplicitity I recommend that this is only done before meals. The module on how to deal with high blood sugars further on in the course covers this aspect.

The insulin calculator is useless and even dangerous if you rely on it for your insulin doses without having done the necessary background study, experiments and personal calculations.

What it does do however is make it unnecessary to do the same calculations every time you are about to eat a meal. Once up and running you only need to change it if any of the three factors need adjusted. The one most frequently changed is the carb sensitivity factors which can vary according to monthly cycles in women and seasonally im most other type ones.

Steven has a portable version on a palm top and we are hoping to put a downloadable version on this site very soon.