Show HN: Glycemic load calculator

[jl2718]

These numbers are completely meaningless for nutrition purposes, and should be generally ignored and replaced with much simpler metrics like CHO digestion half-life.

Glucose levels are increased by digestion/conversion of CHO, and decreased by the effect of insulin. Different types of CHO digest slower(fiber), get into the blood stream faster (glucose) or release more insulin (fructose). A low glycemic load could mean either that all the sugar was immediately deposited in fat cells and now you’re hungry again because it’s dropping and your ghrehlin is spiking, or that your food is still digesting slowly.

Do not be deceived by the low GI/GL. Fructose goes right into your fat cells and makes you hungry by spiking insulin.

Seriously. Ignore GI/GL.

[barftransit]

Fructose does not spike insulin.

[geoffreyhale]

I love your idea, but what's simple about CHO digestion half-life? Certainly not finding more information. Can you link to any resources?

[woodandsteel]

Question: is what you are saying widely accepted by the experts who study human digestion, metabolism, and appetite, or is it a minority view?

On a more anecdotal level, I don't find that eating apples makes me hungry shortly afterwards.

[entropie]

> Glycemic index (GI) > GI is a measure of the effects of a foods carbohydrate component on blood sugar levels. Carbohydrates that break down quickly during digestion and rapidly release glucose into the bloodstream are characterized as having a high GI value, whereas carbohydrates that are broken down slowly and release glucose into the bloodstream in a more gradual manner are characterized as having a low GI value.

> Glycemic load (GL) > Glycemic Load takes into account a foods GI value and a standardized 100g portion size and is calculated as GL = GI x available CHO in a 100g serving / 100. Therefore the GL takes into account the amount of carbohydrate consumed and is a more accurate measure of the impact of a food on blood sugars. As a general rule foods that have a low GL usually have a low GI and those with a medium to high GL value almost always have a very high GI value.

http://www.diogenes-eu.org/GI-Database/Default.htm

[tommaho]

Something I rarely see addressed as an outsider and occasional consumer of this kind of info:

I like bagels - this tool reports a GI / GL of 72 / 42 for 100g. I like peanut butter - 27 and 2.7 for 100g of creamy unsalted.

What is the GI / GL of a peanut-buttered bagel? Is it an average? Can I calculate it?

I don't see much advice around working toward an average glycemic load for a meal versus avoiding specific evils. Is that line of thought flawed?

[mrfusion]

I wish there was something like this to see how foods would affect insulin levels.

I’ve taken to calling insulin the fat storage hormone. Surprisingly glycemic index doesn’t always correspond to insulin level rise.

[rbommisetti]

My father has diabetes and I'm looking to create an app like this. Can you elaborate on your idea?

[DanielBMarkham]

There is. I heard somebody is doing an insulinogenic index.

Sorry, I don't have a link. I've been meaning to look into it but haven't had the time.

[ruc0la]

I was wondering: can you really take out from the equation the bodyweight/fit level/age variables? I'm heavy and also very tall and it seems to me an over simplified service using only a macro sheet of products.

[ucaetano]

No. And even if you take in account those factors, the resulting number is irrelevant on an individual basis due to huge variation between individuals.

[mstudio]

Very interesting! I can't speak to the accuracy of the data, but the overall app is quick, simple and easy to use. One minor UI suggestion: in the live search results, perhaps color code the GI, as you do on the results page. e.g. for "Apple", "GI:38" would be in Green to denote "low" instead of Teal.

[assafmo]

Thanks for the feedback! I think I'll do this. :-)

[assafmo]

Done.

[dzhiurgis]

Reminds my little toy project: https://aminosplore.metacursion.com/

It displays 7 main protein content and lets to compare if you wanna build whole picture.

It's super alpha and barely works (Chrome desktop only, loads entire ~40mb SQLite-webassembly database...)

[assafmo]

Cool!

[goldenkey]

A little off-topic but I recently realized the cultural history of "appetizers." Most of us, myself included, are used to appetizers being a restaurant thing. But I did some digging and it appears there is a valid scientific use. "Whetting one's appetite" is useful for everyone if they want to eat healthier meals. The average person consumes _pounds_ of food daily. If you're like me, it is difficult to eat a large meal even if it is tasty and healthy only if by the size of it. A good meal would include an appetizer to get one acclimated to the larger, less tasty, more work-required meal that is next.

Realizing this helped me adjust all my meals to include some kind of appetizer. Now I eat more healthy food and it isnt as much of a mental struggle.

Hope this realization helps someone else too..

[lolc]

I don't understand your comment. Where I live people generally eat way too much. Me included. It's common understanding here that healthy meals would be smaller but people still eat big meals.

If I strain I can understand your comment to say that appetizers allow us to eat the healthier food that follows. This would require the second course to be more healthy than the first. In my experience this is a rare occurrence.

[magic_beans]

You are a rare case. Most people love eating. Most people do not struggle to eat large amounts food. Most people actively enjoy eating large amounts of food.

No other internet forum on the world is as full of people who hate eating as HN and thinspo-gram.

[contingencies]

Searched for walnuts. Got half a dozen pure walnut records (not walnut cake, etc.), GI was always 24 but GL ranged from 1-4. Source says GL is better indicator than GI. Data is all over the place and publication on which it is based is nearly ten years old. Unsure how to interpret. Inclined to ignore.

PS. If anyone has a serious nutrition background and wants to develop algorithms for dietary value prediction based on customized meals and disparate data sources, we are hiring.

[jdc]

Not sure where to find CHO digestion half-life figures

[magic_beans]

Glycemic load is DRAMATICALLY affected by what you are pairing your foods with. Add fat or fiber and the GL drops. This tool isn't all that helpful.

[mrfusion]

Are you saying I could take a tablespoon of oil after a donut and it will change the glycemic load?

I have been taking a cup of Metamucil after I have a dessert with that same theory in mind. Maybe it’s been helping.

[mygo]

you only designed and implemented for the happy path. when a user submits something that isn’t in your database absolutely nothing happens. a submit button would be a start, as well as some sort of “no results found”. Feedback is important to the user.

[assafmo]

Thanks for your feedback! I'll address this.

[assafmo]

Done. :-)

[cranjice]

So I tried entering the term "pasta" and nothing happened. At first I thought it was broken, but then I entered pizza and there were results. An action button and some status detail (# of results found, etc.) might help the UI.

[assafmo]

Weird, pasta is in the database. What happened to you is probably the slowness of the search. The database is loaded entirely to the client and searches are done in memory.

Right now it's pretty jumpy even on high-end hardware. I need to fix this.

[harigov]

Are you sure this data is correct? It shows that boiled brown rice has a glycemic load higher than white rice.

[assafmo]

The data was taken from http://www.diogenes-eu.org/GI-Database/Default.htm (https://github.com/assafmo/gl-calc/tree/master/src_data)

[magic_beans]

There are a wide variety of rice types, and YES, some white rice varieties have a lower glycemic index than brown rice varieties.